Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Dynamic gene expression patterns during in vivo maturation of mouse hepatoblasts Lee, Sam Chi-Hang 2013

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2013_spring_lee_samchihang.pdf [ 15.29MB ]
Metadata
JSON: 24-1.0073493.json
JSON-LD: 24-1.0073493-ld.json
RDF/XML (Pretty): 24-1.0073493-rdf.xml
RDF/JSON: 24-1.0073493-rdf.json
Turtle: 24-1.0073493-turtle.txt
N-Triples: 24-1.0073493-rdf-ntriples.txt
Original Record: 24-1.0073493-source.json
Full Text
24-1.0073493-fulltext.txt
Citation
24-1.0073493.ris

Full Text

DYNAMIC GENE EXPRESSION PATTERNS DURING IN VIVO MATURATION OF MOUSE HEPATOBLASTS  by SAM CHI-HANG LEE M.Sc., Simon Fraser University, 2006 B.Sc., The University of British Columbia, 2003 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Genetics) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) January 2013  © Sam Chi-Hang Lee, 2013  Abstract Hepatoblasts are bipotential fetal liver cells that differentiate into hepatocytes and cholangiocytes, the parenchymal cell types in the adult liver. While genes that regulate hepatoblast differentiation have been identified, less is known about the underlying regulatory networks. Transcriptome studies using mouse whole fetal livers have been undertaken to decipher those networks. However, as 70-80% of cells in the fetal liver are hematopoietic, the resulting transcriptome data were biased toward genes expressed in blood cells and not maturing hepatoblasts. To address this, I performed a temporal transcriptome analysis using highly purified mouse fetal hepatoblasts. First, I showed that the cell surface molecule DLK1 is specifically expressed on hepatoblasts throughout fetal development. Then, fetal liver DLK1+ cells were isolated on embryonic days 10.5, 12.5, 14.5 and 16.5, and their respective transcriptomes were captured in Tag-seq libraries. The Tag-seq data were specifically enriched for hepatic genes, and liver-enriched transcription factors that were detected at marginal levels in previous fetal liver microarray studies were now readily observed. Furthermore, while genes associated with hepatoblast identity were consistently observed in all four DLK1+ libraries, cholangiocyte genes showed a gradual decrease over time, implicating DLK1 to be repressed during cholangiocyte differentiation. To identify the temporal gene expression profiles underlying hepatoblast maturation, 5449 highly expressed genes in the Tag-seq libraries were subjected to K-means clustering to generate 14 gene groups with distinct temporal expression patterns. Gene Ontology terms were differentially enriched in the gene clusters, suggesting temporally co-regulated genes to share biological functions. Further analysis was performed on HMGA2, a transcription factor with high transcript expression in early hepatoblasts.  ii  Expression of the HMGA2 protein was prominent in nascent hepatoblasts but became restricted to a small subset of hepatoblasts over time. Hmga2 mutant embryos showed a slight reduction of hepatoblasts in the fetal liver, and high Hmga2 was correlated with suppressed expression of the hepatic transcription factor HNF4A in liver cells lines. Overall, this study provides the first genome-wide, temporal analysis of the mouse hepatoblast transcriptome, and the Tag-seq libraries provide a strong foundation for understanding genetic programs that drive in vivo hepatoblast differentiation.  iii  Preface  For genome-wide chromatin immunoprecipitation (ChIP-seq) data in Figure 5-6, I collected  the  starting  cell  material  of  E14.5  DLK1+  cells.  The  chromatin  immunoprecipitation experiment was performed by Rebecca Cullum, and sequencing of the isolated chromatin was performed by the sequencing core at the Michael Smith’s Genome Sciences Centre.  The use of mouse models was approved by the UBC Animal Care and Ethics Committee (protocols: A08-0660 and A08-0763). The use of biohazardous materials and chemicals were approved by the UBC Biosafety Committee.  iv  Table of Contents  Abstract ...............................................................................................................................ii Preface ................................................................................................................................iv Table of Contents ................................................................................................................v List of Tables ....................................................................................................................viii List of Figures.....................................................................................................................ix List of Abbreviations...........................................................................................................x Acknowledgements............................................................................................................xii Dedication.........................................................................................................................xiv Chapter 1: Introduction.....................................................................................................1 1.1 Overview of liver function and significance of functional hepatocytes......................1 1.2 Hepatocyte identity is defined by expression of liver-enriched transcription factors..3 1.3 Generating hepatocyte-like cells by mimicking fetal liver development and ectopically expressing hepatocyte nuclear factors ..............................................................4 1.4 Fetal liver development ............................................................................................5 1.4.1 Temporal overview of hepatic fate induction, hepatoblast specification, and migration.......................................................................................................................6 1.4.2 Temporal overview of genetic events during hepatoblast specification, migration, and survival ..................................................................................................7 1.4.3 Hepatoblast differentiation...............................................................................10 1.4.3.1 Morphological events during hepatoblast differentiation ...........................10 1.4.3.2 Genetic control during hepatoblast differentiation .....................................11 1.4.3.3 Cell surface markers expressed on hepatoblasts.........................................13 1.5 Analysis of the fetal liver transcriptome..................................................................15 1.6 Challenges of fetal liver transcriptome analysis ......................................................17 1.7 Sequencing-based transcriptome analysis techniques: SAGE and Tag-seq ..............19 1.8 Aim of study...........................................................................................................23 Chapter 2: Materials and methods..................................................................................31 2.1 Mouse strains .........................................................................................................31 2.2 Genotyping Hmga2 mutants ...................................................................................31 2.3 Immunofluorescence microscopy............................................................................31 2.4 Cell counting in immunofluorescent images ...........................................................33 2.5 Fetal liver dissociation ............................................................................................33 2.6 Flow cytometry analysis and FACS ........................................................................34 2.7 Magnetic activated cell sorting ...............................................................................35 2.8 Tag-seq library construction ...................................................................................36 2.9 Mapping tags and normalizing libraries ..................................................................36 2.10 K-means clustering ...............................................................................................37 2.11 Gene Ontology analysis ........................................................................................38 2.12 cDNA synthesis and qRT-PCR .............................................................................38 2.13 Lentiviral infection ...............................................................................................39 2.14 Cell culture...........................................................................................................39 Chapter 3: Characterization of DLK1+ cells in the mouse fetal liver.............................43 v  3.1 DLK1+ cells are present in the fetal liver.................................................................43 3.1.1 DLK1 is expressed on hepatic cells in the fetal liver ........................................44 3.1.2 Other fetal liver cell types do not express DLK1 ..............................................46 3.2 Enzymatic dissociation of fetal livers is optimal for FACS isolation of DLK1+ cells 49 3.3 Discussion ..............................................................................................................51 3.3.1 Expression of DLK1 is observed predominantly on hepatoblasts but is also present on non-hepatic cells.........................................................................................51 3.3.2 The method of fetal liver dissociation dictates the number of FACS rounds needed for isolating DLK1+ cells to a purity of 90% or higher .....................................54 Chapter 4: Construction and Analysis of liver Tag-seq libraries...................................67 4.1 Generation of Tag-seq libraries using sorted DLK1+ cells and perfused adult liver tissue...............................................................................................................................67 4.1.1 Stage-specific liver genes are detected in the Tag-seq libraries with expected patterns........................................................................................................................68 4.1.2 Genes expressed in other cell types of the fetal liver are detected at low levels or absent in Tag-seq libraries ...........................................................................................71 4.1.3 Liver transcription factors can be detected in Tag-seq libraries.........................73 4.1.4 Temporal gene expression patterns observed in Tag-seq libraries can be validated by qRT-PCR.................................................................................................77 4.2 Clustering analysis of Tag-seq libraries reveals distinctive gene expression profiles in developing hepatoblasts...................................................................................................78 4.2.1 Temporally distinct gene clusters are enriched for different Gene Ontology terms 81 4.3 Discussion ..............................................................................................................83 4.3.1 Tag-seq libraries accurately captured the changing transcriptome during liver development ................................................................................................................83 4.3.2 Temporal gene clusters may yield novel, stage-specific insights for hepatoblasts 87 Chapter 5: Analysis of Hmga2 expression during hepatogenesis and in liver cell lines ......................................................................................................................................... 105 5.1 Hmga2 is highly expressed among cluster A genes ............................................... 105 5.1.1 HMGA2 is expressed in nascent hepatoblasts ................................................ 106 5.1.2 HMGA2 expression is retained in a subset of maturing hepatoblasts .............. 107 5.2 Hmga2-/- embryos have fewer hepatoblasts ........................................................... 107 5.3 HMGA2 downregulates HNF4A in liver cell lines................................................ 109 5.4 Discussion ............................................................................................................ 111 5.4.1 HMGA2 is dynamically expressed in a subset of fetal liver DLK1+ cells........ 111 5.4.2 HMGA2 may regulate two distinct phases of liver development .................... 113 Chapter 6: Future directions ......................................................................................... 128 6.1 Thesis overview.................................................................................................... 128 6.2 Deep sequencing libraries from fetal liver DLK1+ cells will shed light into the defining characteristic of fetal hepatoblasts ................................................................... 128 6.3 Tag-seq libraries will serve as a useful resource for identifying markers of liver stem and progenitor cells ....................................................................................................... 130  vi  6.4 Applications of Tag-seq libraries toward improving current methods of liver disease treatments...................................................................................................................... 132 References........................................................................................................................ 136 Appendix ......................................................................................................................... 158 Appendix A Cluster A genes and their expression in the five Tag-seq libraries (in tags per million). ........................................................................................................................ 158 Appendix B Cluster B genes and their expression in the five Tag-seq libraries (in tags per million). ........................................................................................................................ 161 Appendix C Cluster C genes and their expression in the five Tag-seq libraries (in tags per million). ........................................................................................................................ 174 Appendix D Cluster D genes and their expression in the five Tag-seq libraries (in tags per million). ........................................................................................................................ 186 Appendix E Cluster E genes and their expression in the five Tag-seq libraries (in tags per million). ........................................................................................................................ 192 Appendix F Cluster F genes and their expression in the five Tag-seq libraries (in tags per million). ........................................................................................................................ 201 Appendix G Cluster G genes and their expression in the five Tag-seq libraries (in tags per million). ........................................................................................................................ 207 Appendix H Cluster H genes and their expression in the five Tag-seq libraries (in tags per million). ........................................................................................................................ 210 Appendix I Cluster I genes and their expression in the five Tag-seq libraries (in tags per million). ........................................................................................................................ 218 Appendix J Cluster J genes and their expression in the five Tag-seq libraries (in tags per million). ........................................................................................................................ 223 Appendix K Cluster K genes and their expression in the five Tag-seq libraries (in tags per million). ........................................................................................................................ 231 Appendix L Cluster L genes and their expression in the five Tag-seq libraries (in tags per million). ........................................................................................................................ 236 Appendix M Cluster M genes and their expression in the five Tag-seq libraries (in tags per million). ........................................................................................................................ 238 Appendix N Cluster N genes and their expression in the five Tag-seq libraries (in tags per million). ........................................................................................................................ 245  vii  List of Tables  Table 1-1 Cell types in the adult liver..................................................................................25 Table 2-1 Primers used in genotyping and qRT-PCR...........................................................41 Table 2-2 Antibodies used for immunofluorescence microscopy and FACS. .......................42 Table 3-1 Percentage of DLK1+ fetal liver cells co-expressing various cell type marker at the indicated timepoints. ...........................................................................................................65 Table 4-1 Summary of Tag-seq library constructed. ............................................................93 Table 4-2 Transcript abundance of genes of interest as detected in Tag-seq libraries. ..........96 Table 4-3 Regulatory genes are distributed among the 14 gene clusters. ............................ 102 Table 4-4 Top enriched GO terms based on clusters with similar temporal patterns........... 103 Table 5-1 Top 10 genes in cluster A as ranked by expression (in TPM) in the E10.5 DLK1+ Tag-seq library.................................................................................................................. 118  viii  List of Figures  Figure 1-1 Architecture of the adult liver.............................................................................26 Figure 1-2 Early phases of hepatogenesis in the mouse embryo...........................................27 Figure 1-3 Bipotential hepatoblast differentiate into hepatocytes and cholangiocytes. .........28 Figure 1-4 Tag-seq methodology.........................................................................................29 Figure 3-1 DLK1+ cells are present throughout fetal liver development...............................58 Figure 3-2 Expression of HNF4A and DLK1 in the E9.5 liver bud. .....................................59 Figure 3-3 Expression of DLK1 and HNF4A largely co-localizes during fetal liver development........................................................................................................................60 Figure 3-4 DLK1 and CD45 are expressed in mutually exclusive populations. ....................61 Figure 3-5 DLK1 does not overlap with PECAM expression in the fetal liver. ....................62 Figure 3-6 DLK1 and desmin expression slightly overlaps during fetal liver development. .64 Figure 3-7 Dissociation methods for isolating DLK1+ fetal liver cells. ................................66 Figure 4-1 Tag-seq libraries faithfully capture expression of stage-specific liver genes........95 Figure 4-2 Temporal expression patterns of hepatoblast/cholangiocyte TFs in Tag-seq libraries...............................................................................................................................97 Figure 4-3 qRT-PCR validation of select signaling molecules with no characterized roles in hepatoblast differentiation. ..................................................................................................99 Figure 4-4 Fourteen temporal clusters were generated from K-means clustering................ 100 Figure 4-5 Genes with preferential expression in early hepatoblasts or late hepatocytes are enriched for specific Gene Ontology biological processes. ................................................ 104 Figure 5-1 HMGA2 is expressed in E9 and E9.5 hepatoblasts. .......................................... 120 Figure 5-2 A subset of DLK1+ hepatoblasts also express HMGA2 in the E9.5 liver bud and E10.5 liver. ....................................................................................................................... 121 Figure 5-3 Few DLK1+ co-express HMGA2 in E12.5 and E14.5 fetal liver. ...................... 122 Figure 5-4 Fewer fetal liver DLK1+ cells are present in Hmga2-/- embryos. ....................... 123 Figure 5-5 Hmga2 mutant fetal hepatoblasts retain DLK1 expression................................ 124 Figure 5-6 HNF4A occupies several sites within Hmga2 in E14.5 DLK1+ hepatoblasts..... 125 Figure 5-7 Hmga2 expression in HPPL cells. .................................................................... 126 Figure 5-8 Expression level of HMGA2 and HNF4A are inversely correlated in HepG2 cells. ......................................................................................................................................... 127  ix  List of Abbreviations BMP = bone morphogenetic protein BSA = bovine serum albumin ChIP-seq = chromatin immunoprecipitation sequencing DABCO = 1,4-diazabicyclo[2.2.2]octane Dex = dexamethasone DLK1 = delta-like 1 homolog DMEM = Dulbecco’s minimal essential media E = embryonic day ECM = extracellular matrix EGF = epidermal growth factor EHS = Engelbreth-Holm-Swarm ITS = insulin/transferrin/selenium FACS = fluorescence activated cell sorting FBS = fetal bovine serum FGF = fibroblast growth factor HCC = hepatocellular carcinoma HMGA2 = high mobility group AT-hook 2 HNF = hepatocyte nuclear factor HPPL = hepatic progenitors proliferating on laminin LYVE1 = lymphatic vessel endothelial hyaluronan receptor 1 MACS = magnetic activated cell sorting NGS = next generation sequencing  x  OSM = oncostatin M PECAM = platelet endothelial cell adhesion molecule Tag-seq = tag sequencing TF = transcription factors TFBS = transcription factor binding sites TPM = tags per million TSS = transcriptional start site PFA = paraformaldehyde SAGE = serial analysis of gene expression SD = standard deviation SMA = smooth muscle actin SSOOHE = sans singletons and one-offs of highly expressed tags STM = septum transversum mesenchyme qRT-PCR = quantitative reverse transcription polymerase chain reaction  xi  Acknowledgements  I offer my enduring gratitude to the faculty, staff and my fellow students at the UBC, BCCA Cancer Research Centre, and Genome Sciences Centre, who have inspired me with your passion and dedication to science. Thanks to my Committee Members Dr. Keith Humphries, Dr. Aly Karsan, and Dr. Rob Kay for keeping me on track during the course of this PhD study. I owe particular thanks to my supervisor Dr. Pamela Hoodless. Her penetrating questions taught me not to assume anything, and I am forever indebted for the patience and freedom she has given me to pursue teaching and outreach opportunities outside the lab.  I thank all the current and past members of the Hoodless lab for making it enjoyable to come to work everyday, especially: Kristie Baas for helping with administrative issues, big and small; Rebecca Cullum for being an unflappable presence in the lab; Victoria Garside for her unlimited supply of finger-pointing nitrile gloves; Dr. Olivia Alder and Dr. Wei Wei for their scientific prowess in taking the liver project to the next level; Dr. Juan Hou for her scientific mentoring, particular during my first months in the lab; Dr. Pavle Vrljicak for his valuable advice of start writing early; and all other unnamed members who have shared laughter and their passion for science and life with me. I would also like to thank members of the Flow Core Facility for their expertise and patience, which were crucial in bringing this thesis study to fruition. Heartfelt thanks also go out to Michael Copley for engaging scientific discussions and generous donation of reagents; Dr. Alex Chang for discussions on life after graduate school; and Dr. Gordon Robertson for embodying the true interdisciplinary scientist.  xii  Special thanks are owed to my family, who has supported me throughout my years of education, both morally and financially. To all my friends who may not understand this body of work but nonetheless had unwavering faith in me to produce something beneficial to mankind, this is the final product of my hard work in the past six years. Last but definitely not least, lots of thanks and love to Malina Leung for urging me to complete the final and toughest stages of this dissertation.  xiii  Dedication  For my Brother, Nick, and the quest toward building an organ.  xiv  Chapter 1: Introduction  1.1  Overview of liver function and significance of functional hepatocytes The liver is an essential organ for maintaining mammalian homeostasis. Some of the  liver’s essential functions include blood toxin removal, bile secretion, and lipid and glycogen storage. The functional cellular units of a liver are hepatocytes, which are arranged as repeating, stereotypic hexagons in the mature organ (Si-Tayeb et al., 2010; Spear et al., 2006; Figure 1-1). At the corner of each hexagon is the portal triad, which consists of the hepatic artery and portal vein (blood vessels coming from the heart and gut, respectively) as well as the bile duct (transport bile synthesized by the hepatocytes to the gall bladder). In the middle of the hexagon is the central vein, which takes blood back to the heart. The hepatocytes, arranged in cords, radiate outward from the central vein towards the edges of the hexagon. These hepatic cords are lined by sinusoidal endothelial cells, which mediate the exchange of metabolites between hepatocytes and the incoming blood. Additional cell types are also present in the adult liver (Table 1-1). Cholangiocytes make up the bile ducts that are responsible for bile transport, whereas pit cells and Kupffer cells cooperate to mount an immune response during localized assaults such as hepatitis (Wisse et al., 1997). Stellate cells are fibroblast-like cells that store vitamin A and control the deposition of extracellular matrix components in the sinusoids (Geerts, 2001). Lastly, oval cells, proposed to be stem cells of the adult liver, are needed during liver regeneration to generate new hepatocytes after massive tissue loss or necrosis (Fausto, 2000). Severe injuries to the liver can lead to chronic or acute hepatic failure. Alcohol abuse and hepatitis are common causes of chronic cirrhosis, whereas acute hepatic failure is defined  1  as rapid onset liver failure that happens without a history of liver disease (Heidelbaugh and Bruderly, 2006; O’Grady et al., 1993). Another cause of hepatic failure is end-stage liver cancer, and hepatocarcinoma has been an increasing threat to Canadians in recent years. According to statistics from the Canadian Cancer Society, 2000 new cases of liver cancer were diagnosed in 2012, resulting in the death of 900 patients. Significantly, the mortality rate of patients living with liver cancer is 2.2%, making it the only common cancer with an increasing mortality rate from 1998 to 2007. Therefore, much research is needed to improve current methods for treating patients living with liver cancer and other forms of liver disease. While the liver can regenerate itself after up to 70% hepatectomy, the regeneration process is slow and thus insufficient for repopulating a rapidly failing liver. Instead, liver transplantation is the most effective treatment for patients with end-stage liver failure. However, the demand for organ donors is greater than the supply, as highlighted by the 50% mortality rate of acute hepatic failure patients (de Rave et al., 2002). In Canada, close to 1000 patients await a liver graft per year (Shah et al., 2006) and, in the province of British Columbia, the mortality rate for patients awaiting a liver transplant is 30% (Haque et al., 2010). As a result, alternative treatment methods such as cell therapy and bioartificial livers are being intensively pursued (Strain and Neuberger, 2002; Vosough et al., 2011). However, in order for these alternative methods to be successful, a reliable method of generating hepatocytes under laboratory conditions is first needed.  2  1.2  Hepatocyte identity is defined by expression of liver-enriched transcription  factors A defining characteristic of fetal and adult hepatocytes is their abilities to synthesize essential blood serum proteins such as albumin, transthyretin, and alpha 1-antitrypsin (Clissold and Bishop, 1981; Dickson et al., 1985; Freeman et al., 1981). Early experiments seeking to identify liver-enriched proteins that bind to the promoters of those liver genes led to the discovery of FoxA, HNF4, and CEBP (Costa et al., 1989; Friedman et al., 1989). While these transcription factors bind DNA with high affinity and sequence specificity, they are structurally distinct from each other. The group of FoxA proteins, originally classified as hepatocyte nuclear factor 3 (HNF3), consists of three members (FoxA1, 2, and 3) that contain the winged-helix/forkhead DNA-binding domain (Costa et al., 1989; Kaestner et al., 2000). HNF4, currently known as hepatocyte nuclear factor 4A (HNF4A), is an orphan nuclear receptor in the steroid hormone receptor superfamily (Sladek et al. 1990). Lastly, the family of CCAAT/enhancer binding proteins has two members (CEBP alpha and beta) that are characterized by a basic region involved in DNA binding and a leucine zipper motif that is required for protein dimerization (Lamb and McKnight, 1991). Subsequently, these transcription factors were found to be also expressed in fetal hepatic progenitors (Duncan et al., 1994; Monaghan et al., 1993; Westmacott et al., 2006), and they act on key aspects of hepatogenesis such as expression of the first liver marker genes (Lee et al., 2005) and differentiation of the hepatic progenitors (Battle et al., 2006; Yamasaki et al., 2006). Further research identified HNF1 and HNF6 – transcription factors containing the Pou and onecut homeodomain, respectively – to also bind the promoters of hepatocyte genes and regulate their expression (Lannoy et al., 2002; Rollier et al., 1993). Not  3  surprisingly, similar to other liver-enriched transcription factors, both HNF1 and HNF6 have roles in regulating differentiation of fetal liver hepatoblasts (Clotman et al., 2002; PlumbRudewiez et al., 2004; Yamasaki et al., 2006). Taken together, these studies effectively highlight the importance of liver-enriched transcription factors during both the adult and embryonic stages of liver development.  1.3  Generating hepatocyte-like cells by mimicking fetal liver development and  ectopically expressing hepatocyte nuclear factors Methods to generate hepatocytes in vitro have been developed (Baxter et al., 2010). Many such protocols subject embryonic stem cells or adult multipotent progenitors through a multi-step procedure where the cells are first induced into endoderm, then hepatic progenitors or hepatoblasts, and finally hepatocyte-like cells (Soto-Gutierrez et al., 2007; Roelandt et al., 2010). The hepatoblast induction step involves the generation of a fetal liverlike environment; specifically, the stem or progenitor cells are co-cultured with hepatic endothelial and stellate cells lines along with the addition of hepatocyte (HGF) and fibroblast (FGF) growth factors in the growth media. In the midgestation embryo, HGF is required for the survival of hepatoblasts as well as the formation of cell-cell junctions with other hepatoblasts (Schmidt et al., 1995), while FGFs signal to the hepatic precursors to initiate expression of early liver genes and expansion of the nascent liver bud (Jung et al., 1999). As the maintenance of stem and progenitor cells is both time- and labour-intensive, researchers have recently begun to explore generating hepatocyte-like cells from differentiated cells. For instance, hematopoietic cells expressing the oncostatin M ! receptor showed signs of hepatic transdifferentiation upon forced expression of the hepatic  4  transcription factor Hnf4! (Khurana et al., 2010). More recently, fibroblasts were used to generate functional hepatocyte-like cells, either through ectopic expression of Hnf4! in combination with one of the three FoxA transcription factors in embryonic or adult fibroblasts (Sekiya and Suzuki, 2011), or forced expression of Gata4, Hnf1a, and Foxa3 in adult mouse tail tip fibroblasts (Huang et al., 2011). Intriguingly, the transcription factors involved in reprogramming fibroblasts also play major roles during normal hepatogenesis. Gata4 and FoxA affect the early phases of liver specification by potentiating expression of early liver genes in the hepatic foregut endoderm (Bossard and Zaret, 1998; Friedman and Kaestner, 2006; Gualdi et al., 1996; Lee et al., 2005). On the other hand, Hnf4! and Hnf1! act during the hepatoblast differentiation stage to activate genes that establish hepatocyte identity (Li et al., 2000; Parviz et al., 2003). Taken together, factors that dictate fetal hepatogenesis are also potent drivers of hepatocyte transdifferentiation in vitro. However, current transdifferentiation protocols still suffer from poor efficiency. For instance, the transdifferentiated cells do not express the full arsenal of metabolic enzymes as in hepatocytes, and they are also inferior to primary hepatocytes in their ability to repopulate an injured liver (Huang et al., 2011; Sekiya and Suzuki, 2011). Thus, a better understanding of normal hepatogenesis would likely provide insights for optimizing current strategies of generating hepatocyte-like cells.  1.4  Fetal liver development Hepatogenesis is a multi-step process involving the interplay of various cell types and  signaling molecules. The next sections will describe the key cellular and morphological  5  events during hepatogenesis, followed by an overview of the genetic components that drive these cellular changes.  1.4.1  Temporal overview of hepatic fate induction, hepatoblast specification, and  migration Development of the fetal liver has been extensively characterized (reviewed by Zaret, 2002; Zaret et al., 2008; and Crawford et al., 2010). In the mouse embryo, hepatic precursors first appear around embryonic day (E) 8.5, when a portion of the ventral foregut endoderm adopts the hepatic fate and begins to express albumin and alpha-fetoprotein, the first markers of liver specification (Cascio and Zaret, 1991; Schmid and Schulz, 1990; Figure 1-2 top panel). A day later, the nascent hepatoblasts then delaminate from the endoderm monolayer and invade the septum transversum mesenchyme (STM), which is situated between the hepatoblasts and the cardiac mesoderm. This invasion can be clearly visualized as HNF4Apositive  hepatoblasts  spread  and  become  interspersed  between  GATA4-positive  mesenchymal cells (Zhao and Duncan, 2005; Figure 1-2 bottom panel). By E10.5, sufficient hepatoblasts have ingressed into the surrounding mesenchyme to form the nascent liver bud (Theiler, 1989). At this stage, PECAM-positive endothelial cells begin to line the hepatoblasts and form immature hepatic sinusoids (Matsumoto et al., 2001; Crawford et al., 2010). Concurrently, hematopoietic progenitors from the yolk sac begin homing into the fetal liver (Sasaki and Matsumura, 1986). In the embryo, primitive hematopoiesis is first observed in the yolk sac. Hepatoblasts express many blood serum proteins (such as albumin and apoliproproteins) that are also produced in the yolk sac, making the fetal liver an ideal organ for supporting fetal hematopoiesis after midgestation (Isern et al., 2008; Meehan et al. 1984).  6  By E11.5, the liver has increased greatly in size due to the expansion of the hepatic and hematopoietic compartments (Crawford et al., 2010). At E12.5, the overall cellular composition and architecture of the fetal liver remains unchanged (Crawford et al., 2010). However, by E13.5, fissures divide the fetal liver into four lobes, and expression of cholangiocyte and hepatocyte genes begin, marking the beginning of hepatoblast differentiation (Clotman et al., 2002; Germain et al., 1988; Lemaigre, 2003; Shiojiri 1981, 1984). Hematopoietic cells continue their expansion in the E13.5 fetal liver, and the peak of fetal liver hematopoiesis is reached by E14.5 (Kurata et al., 1998; Rich and Kubanek, 1979), where close to 80% of fetal liver cells are hematopoietic (Sigal et al., 1994).  1.4.2  Temporal overview of genetic events during hepatoblast specification,  migration, and survival Induction of early liver markers in E8.5 hepatic progenitors is intricately modulated. Hepatic progenitors, derived from the foregut endoderm, interact with adjacent cell populations of mesodermal origin to initiate expression of the hepatic markers (Figure 1-2). In ex vivo explant studies, hepatic fate of the foregut endoderm could only be induced when either the cardiac mesoderm or septum transversum mesenchyme was present (Gualdi et al., 1996; Jung et al., 1999; Rossi et al., 2001). The factors triggering this induction were found to be FGF1, FGF2 (both secreted by the cardiac mesoderm), and BMP4 (secreted by the septum transversum mesenchyme), as addition of these factors in lieu of the mesenchyme tissues could similarly induce cultured endoderm to adopt hepatic fate.  7  Intrinsic factors in hepatoblasts also regulate the expression of early liver markers. The FOXA transcription factors regulate a wide range of developmental processes, and recently have been shown to act as pioneering factors that govern gene expression through chromatin remodeling (Friedman and Kaestner, 2006; Zaret et al., 2008). Before hepatic induction, the foregut endoderm still expresses GATA4 (Watt et al. 2007), and GATA4 partners with FOXA transcription factors in the pre-hepatic endoderm to loosen the chromatin structure around Afp and Alb, leading to their subsequent expression in nascent hepatoblasts (Bossard and Zaret, 1998; Crowe et al., 1999; Gualdi et al., 1996). The importance of the FOXA proteins in hepatic gene induction is further exemplified in mutant studies, as embryos lacking Foxa1 and Foxa2 do not turn on the hepatic program even in the presence of the FGF and BMP inductive signals (Lee et al., 2005). Once the hepatoblasts are specified, they enter the next phase of hepatogenesis by migrating into the STM, forming a liver bud (Zaret, 1998). During this process, cell adhesion and extracellular matrix (ECM) proteins play important roles. Support for this claim was first observed when ES cells lacking functional integrin !1 were unable to colonize the liver in chimeric animals (Fässler and Meyer, 1995). Subsequent studies with mutant mice revealed that levels of other cell adhesion and ECM proteins such as E-cadherin, laminin, and collagen, are also regulated during liver bud migration. Hepatoblasts invade into the septum transversum mesenchyme as an epithelial sheet (Wilson et al., 1963). Hex encodes the hematopoietically expressed homeobox protein, and is essential for embryonic development of the brain, liver, and thyroid (Martinez Barbera et al., 2000). In Hex mutants, hepatoblasts fail to form a pseudostratified epithelium, and the lack of laminin degradation between hepatoblasts and the encasing mesenchyme results in a drastically reduced liver bud (Bort et  8  al., 2006). Another homeobox protein that controls liver bud expansion is Prox1, which regulates retina, ear, and liver development (Bermingham-McDonogh et al., 2006; Dudas et al., 2004; Dyer et al., 2003). In Prox1 mutants, hepatoblasts fail to populate the septum transversum mesenchyme due to an inability to downregulate E-cadherin and break down the surrounding collagen IV (Sosa-Pineda et al., 2000). Similar to early liver gene expression, non-hepatic cells are crucial for hepatic bud expansion. Ex vivo explant experiments established a role for FGF8 secreted by the cardiac mesoderm in inducing liver bud outgrowth (Gualdi et al., 1996; Jung et al. 1999). Similar explant experiments showed that the STM also induces liver bud outgrowth by the secretion of BMP4 (Rossi et al., 2001). Interestingly, endothelial cells are also required for invasion into the STM. Not only do Flk1 mutant embryos produce no endothelial cells, their hepatoblasts also show no migration into the STM (Matsumoto et al., 2001). Taken together, hepatoblast migration is controlled by signals both intrinsic and extrinsic to the invading hepatoblasts. After hepatoblasts complete their invasion into the STM, they enter the proliferation phase of hepatogenesis in order to support hematopoiesis (Kamiya et al., 1999) and overall fetus growth (Schmidt et al., 1995). Not surprisingly, hepatoblasts rely on cell autonomous as well as non-autonomous signals for their proliferation. For instance, mutants of Hlx, a homeobox protein expressed in the fetal liver mesenchyme (Lints et al., 1996), exhibit severe hypoplasia of the fetal liver (Hentsch et al., 1996). As the early liver marker Afp was properly expressed in mutant hepatoblasts, the cell non-autonomous signal from the mesenchymal cells is not involved in hepatic specification but instead supply a pro-hepatogenic signal after specification. In addition, hepatoblasts express components of the hepatocyte growth factor  9  (HGF) and Wnt signaling pathways to mediate survival cues (Decaens et al., 2008; Schmidt et al., 1995; Tan et al., 2008). The transcription factors Foxm1 and cJun are also involved in maintaining hepatoblast survival, as mutant hepatoblasts fail to enter mitosis or have increased apoptosis, respectively (Eferl et al., 1999; Krupczak-Hollis et al., 2004). More recent studies using primary hepatoblasts have also identified transcription factors with pro(Gli1, Prox1) or anti-mitotic (Tim2) roles (Hirose et al., 2009; Kamiya et al., 2008; Watanabe et al., 2007). Interestingly, genes that do not directly regulate mitosis can also affect hepatoblast survival. Smad2 and Smad3 encode intracellular signaling proteins of the TGF! signaling pathway, and Smad2+/-; Smad3+/- embryos develop hypoplastic livers due to immature cell-cell contacts in mutant hepatoblasts (Weinstein et al., 2001).  1.4.3  Hepatoblast differentiation The next step in fetal liver development is hepatoblast differentiation, where  bipotential hepatoblasts begin committing toward either the hepatocyte or cholangiocyte lineage. This process occurs around E13.5, when hepatocyte and cholangiocyte markers begin to be expressed on hepatoblasts (Germain et al., 1988). In addition, morphological changes are associated with hepatoblast differentiation.  1.4.3.1  Morphological events during hepatoblast differentiation Several remodeling events occurring in the late-gestation fetal liver are essential in  priming the fetal liver for its postnatal metabolic role. Developing hepatocytes increase in number, cell size, and cell-cell contacts with neighbouring hepatocytes, resulting in polarized hepatocytes and tight hepatic cords (Battle et al., 2006; Greengard et al., 1972; Parviz et al.,  10  2003; Figure 1-3). Along with hepatic cord formation, these cells also begin acquiring apicalbasal polarity, and their sinusoid structures with the surrounding vasculature continue to mature (Parviz et al., 2003; Si-Tayeb et al., 2010; Zhao and Duncan, 2005). Cholangiocytic hepatoblasts also undergo cellular remodeling. Cells closest to the portal vein re-arrange into a monolayer to form rudimentary bile ducts, which eventually pinch off to form ductal plates and subsequently the mature bile ducts (Antoniou et al., 2009; Lemaigre, 2003; Si-Tayeb et al., 2010; Zorn, 2008; Figure 1-3).  1.4.3.2  Genetic control during hepatoblast differentiation The main driver of hepatocyte differentiation is the Hnf4! gene, and its effects on  hepatocyte differentiation have been extensively studied in mutant models (Li et al., 2000; Parviz et al., 2003; Battle et al. 2006). While conditional Hnf4!-/- hepatoblasts have normal morphology at E12.5, the expression of many blood serum proteins and hepatocyte transcription factors are either completely absent or greatly diminished (Li et al., 2000). By E18.5, the expression of approximately 30 cell junction proteins is also affected, resulting in a disorganized hepatic epithelium (Parviz et al., 2003; Battle et al., 2006). HNF4A was subsequently found to bind promoters of many of the affected cell adhesion genes, suggesting the transcription factor to control hepatocyte differentiation via transcriptional regulation (Battle et al., 2006). In cholangiocyte differentiation, Hnf1" and Hnf6 are the central effectors within an intricate regulatory cascade. Both Hnf6-/- and Hnf1b-/- mutant embryos exhibit disorganized bile ducts due to the production of hybrid hepatoblasts expressing markers for both hepatocytes and cholangiocytes (Coffinier et al., 2002; Clotman et al., 2002). Hepatoblasts  11  from Hex or Foxm1 mutants also show defects in cholangiocyte differentiation (Hunter et al., 2007; Krupczak-Hollis et al., 2004). These mutant fetal livers express lower levels of Hnf6 and Hnf1", suggesting Hex and Foxm1 to act further upstream in the cholangiocyte differentiation regulatory cascade. In addition to positive regulators of Hnf1" and Hnf6, negative regulators such as Cebp! and Tbx3 have also been found, and these mutants upregulate Hnf1" and Hnf6 in conjunction with an excess of biliary cells (Yamasaki et al., 2006; Suzuki et al., 2008; Lüdtke et al., 2009). Not surprisingly, genes affecting Cebp! expression can also regulate cholangiocyte differentiation: Ctnnb1-/- and Sox9-/- hepatoblasts express lower levels of Cebp!, which delayed the timing of their cholangiocyte lineage commitment (Antoniou et al., 2009; Tan et al., 2008). Other transcription factors such as Sall4 and Tim2 can also control cholangiocyte differentiation in cultured hepatoblasts (Oikawa et al., 2009; Watanabe et al., 2007), but their effects on hepatoblasts in vivo remain to be explored. In addition to transcription factors, signaling pathways have also been shown to direct cholangiocyte differentiation. In Hnf6-/- mutants, the expression of TgfbrII, a receptor for TGF! ligands, is upregulated (Plumb-Rudewiez et al., 2004), whereas expression of 2macroglobulin and follistatin, both TGF! antagonists, is reduced (Clotman et al., 2005). Furthermore, the domain of active TGF! signaling in Hnf6-/- mutant embryos is expanded beyond the periportal region, which leads to an excess of hepatoblasts with bile duct marker expression (Clotman et al., 2005). The Notch signaling pathway has also been shown to promote cholangiocyte differentiation in developing hepatoblasts. In the fetal liver, bile ducts cells express high levels of the JAGGED1 ligand, while the neighbouring hepatoblasts express the NOTCH2 receptor (Tanimizu and Miyajima, 2004). Interestingly, the expression 12  patterns of these two proteins are misregulated in Cebp!-/- fetal livers, a phenotype that could underlie the observed overproduction of cholangiocytes in the mutants (Yamasaki et al., 2006). The role for Notch signaling in cholangiocyte differentiation was confirmed via in vitro studies, as primary hepatoblasts with activated Notch signaling repress the hepatocyte program and in contrast upregulate cholangiocyte genes (Tanimizu and Miyajima, 2004).  1.4.3.3  Cell surface markers expressed on hepatoblasts As described above, cell surface receptors play important roles during hepatoblast  differentiation. In fact, the presence or absence of select cell surface markers has been used to characterize and isolate hepatoblasts from the fetal liver. Collectively, liver cells with hepatic potential do not express markers of the hematopoietic lineage, specifically leukocyte common antigen/CD45, lymphocyte antigen 76/Ter119, and kit oncogene/cKit (Minguet et al., 2003; Suzuki et al., 2000; Suzuki et al., 2002). In contrast, cell adhesion molecules such as E-cadherin and the integrin alpha 6 and beta 1 subunits were specifically expressed on hepatic cells (Nierhoff et al., 2005; Suzuki et al., 2000). Moreover, signaling molecules such as MET, the HGF receptor; and GPC3, the proteoglycan glypican-3, have also been used to mark hepatic progenitors in mice and rats, respectively (Grozdanov et al., 2006; Suzuki et al., 2002). Dlk1, or delta-like 1 homolog, encodes a transmembrane protein with six EGF-like repeats that is related to the Notch/Delta family of membrane-bound signaling molecules (Smas and Sul, 1993; Lai, 2004). Originally identified as fetal antigen 1 (FA1) or preadipocyte factor 1 (PREF-1), DLK1 has a shortened intracellular tail and lacks an important region of the extracellular Delta/Serrate/LAG-2 domain (Sul, 2009), and thus  13  likely functions as an atypical Delta ligand (Bray et al., 2008). In mammals, DLK1 can be detected in the embryonic liver, lung, limb bud, vertebrae, and pancreas, but post-natal expression is restricted to neuroendocrine cell types such as the pituitary, adrenal gland, islet, as well as preadipocytes (Carlsson et al., 1997; Floridon et al., 2000; Smas and Sul, 1993; Tanimizu et al., 2003; Wang et al., 2006). As supported by the wide range of tissues expressing Dlk1, the gene is involved in a diverse array of developmental processes, such as adipogenesis (Lee et al., 2003; Moon et al., 2002; Wang et al., 2006), skeletal development (Miyaoka et al., 2010; Waddell et al., 2010), hematopoiesis (Sakajiri et al., 2005; Wu et al., 2008), and even cancer progression (Yanai et al., 2010). Notably, Dlk1 was activated in regenerating rat livers, and the expression was confined to oval cells, the putative hepatic progenitor cells in the adult liver (Lemire et al., 1991). Interestingly, while the Notch signaling pathway can regulate hepatoblast differentiation (Ader et al., 2006; Tanimizu and Miyajima; 2004), no defects in liver development have been documented in Dlk1 mutant or transgenic animals (Lee et al., 2003; Moon et al., 2002). In a screen for proteins that are enriched in fetal hepatic cells, DLK1 was identified as a top hit among cell surface molecules (Tanimizu et al., 2003). Using an antibody specific to DLK1, E14.5 DLK1+ fetal liver cells were isolated. By qPCR, the expression of hepatic serum proteins (Alb and Afp) and transcription factors (Hnf1b, Hnf4a, and Foxa2) was detected in DLK1+ cells, while genes expressed in mature hepatocytes (Tat and Cps) and cholangiocytes (Krt19) were notably absent. Interestingly, both hepatocyte and cholangiocyte genes became activated when DLK1+ cells were cultured in vitro. When Notch signaling was suppressed in DLK1+ primary cultures to promote hepatocyte differentiation, the resulting cells exhibited condensed cytosol and rounded nuclei, a morphology that resemble  14  differentiated hepatocytes (Tanimizu and Miyajima, 2004). Furthermore, E14.5 DLK1+ cells are capable of homing to and repopulate an injured adult liver (Tanimizu et al., 2003). These preliminary results suggest E14.5 fetal liver DLK1+ cells to have full potential for hepatocyte differentiation. To further study their differentiation ability, E14.5 DLK1+ cells were maintained in long-term cultures to establish a cell line termed hepatic progenitor cell proliferating on laminin, or HPPL (Tanimizu et al., 2004). When HPPL cells were grown on an EngelbrethHolm-Swarm (EHS) gel substrate and in the presence of oncostatin M, HPPL cells acquired additional hepatocyte traits such as polysaccharide storage and ammonium removal. Significantly, HPPL cells could also differentiate into cholangiocytes. When cultured on collagen gels, they formed tube-like structures resembling bile ducts (Tanimizu et al., 2004). With the application of a hybrid Matrigel-extracellular matrix sandwich culturing system, HPPL cells become polarized and developed into ductal plates that are complete with a basement membrane and a functional solute secretion system (Tanimizu et al., 2007; Tanimizu et al., 2009). Taken together, these findings strongly suggest DLK1+ cells to be bipotential progenitor cells in the midgestation fetal liver.  1.5  Analysis of the fetal liver transcriptome As described above, the differentiation and maturation of hepatoblasts, hepatocytes,  and cholangiocytes require the coordinated expression of many transcription factors and signaling molecules. To investigate the underlying mechanisms that drive these processes, various groups have performed large-scale gene expression analysis using liver cells from either mutant mouse models or immortalized cell lines (Ader et al., 2006; Gresh et al., 2005;  15  Papoutsi et al., 2007). While these systems provide convenient platforms for studying hepatoblast differentiation, they may deviate sufficiently from hepatogenesis in utero and thereby limiting the applicability of their findings. Thus, it is essential to analyze the transcriptome of normal liver cells as they differentiate and mature during the course of embryogenesis. Several groups have performed transcriptome analysis of the fetal liver over the course of development. Jochheim et al. used microarrays to capture the transcriptomes of pre-hepatic tissue (E7.5 whole embryo), fetal livers (E11.5 and E13.5), and adult livers. The majority of differentially expressed genes from this analysis were chemokines, proteases, and other metabolic enzymes that reflect roles in mature, adult hepatocytes (Jochheim et al., 2003). Immature hepatocytes in the fetus progressively differentiate toward functional hepatocytes in the adult, which is reflected by an increase in the expression of genes encoding serum proteins such as Afp and Alb. However, the expression of hepatic genes showed either a sharp decrease at E11.5 (Cebp" and Hnf4!) or was below the detection threshold (Met). A second study used qRT-PCR to detect expression of eight liver-enriched transcription factors (Hnf1!, Hnf1", Hnf3!, Hnf4!, Hnf6, Foxa2, Cebp!, and Cebp") during fetal liver development (Jochheim et al., 2004). With the exception of Cebp!, the expression of all transcription factors was drastically repressed between the E11.5 and E13.5 fetal livers, which coincides with the peak of hematopoiesis in the fetal liver (Kurata et al., 1998; Rich and Kubanek, 1979). A follow-up study using microarrays found over 100 genes to be significantly upregulated in E11.5 and E13.5 livers relative to the pre-hepatic embryo. However, most of these genes were not associated with hepatic differentiation but instead with hematopoietic expansion (Jochheim-Richter et al., 2006). A separate group that 16  analyzed the transcriptomes of E10.5-E16.5 fetal, normal adult, and regenerating liver had a similar finding, whereby genes upregulated in the fetal liver were also related to proliferation of hematopoietic cells (Otu et al., 2007). To date, the most complete temporal profiling of the fetal liver transcriptome was done by Li et al., where they used microarrays to capture the hepatic transcriptome daily beginning at E11.5, as well as post-natal timepoints on Days 3, 7, 14, and 21 (Li et al., 2009). Transcription factors that regulate cell cycle and DNA replication were found to be highly expressed in the fetal stages, whereas hepatic transcription factors were only detected at high levels postnatally. A total of 8460 genes showing differentially expression across the timepoints were analyzed by hierarchical clustering, and four classes of temporal expression patterns were observed. Two of the four classes showed high gene expression levels in the fetal liver, and Gene Ontology analysis showed the two groups to be enriched for genes associated with ‘blood coagulation’ and ‘heme biosynthesis’. To gain insight into the underlying transcriptional networks, the most differentially expressed genes were subjected to promoter analysis to try and identify cis-regulatory motifs that are enriched during liver development. Specifically, the enriched motifs shift from hematopoietic transcription factors (such as SP1, Mef2, Evi1, and E2f1) in the fetal stages, to hepatic transcription factors (Hnf4a and Smad3) in the postnatal and adult stages.  1.6  Challenges of fetal liver transcriptome analysis Evidently, while previous transcriptome studies using whole fetal livers were useful  in highlighting the liver’s transition from being mainly hematopoietic in the fetus to metabolic in the adult, they did not generate new insights surrounding fetal hepatoblast  17  differentiation. Specifically, the expression of key transcription factors that drive fetal hepatoblast differentiation could not be robustly detected. This can be attributed largely to the cellular composition of the fetal liver. The organ is the main site of hematopoiesis from E12.5 to E14.5, where the hematopoietic compartment makes up 70-80% of total fetal liver cells (Crawford et al., 2010; Sigal et al., 1994). In contrast, hepatoblasts constitute only 210% of cells in E12.5/E14.5 fetal livers (Nierhoff et al., 2005; Tanimizu et al., 2003; Sigal et al., 1994; Suzuki et al., 2008). Therefore, when whole fetal livers are used for gene expression studies, only a small fraction of the transcriptome data would be derived from hepatoblasts. This is clearly exemplified by transcriptome libraries generated using whole fetal livers and serial analysis of gene expression (SAGE, described in detail below). In the E12.5 fetal liver SAGE library, transcripts derived from Hbb-y, a gene encoding the beta-like embryonic chain of hemoglobin, were abundantly detected at close to 10,000 counts. In contrast, transcripts derived from hepatic genes were detected at significantly lower levels. Specifically, expression of the secreted serum protein, albumin, was detected at just over 300 counts, and the hepatic transcription factors Hnf4! and Foxa2 were detected at counts of two and  zero,  respectively  (unpublished  results,  Hoodless  lab:  http://www.mouseatlas.org/data/mouse/libraries/SM001). In the E18.5 fetal liver library, expression of albumin is more readily detectable at 5169 counts, which is reflective of the increased number of hepatocytes in the perinatal fetal liver. However, both Foxa2 and Hnf4! expression were still detected at the low level of a single count (unpublished results, Hoodless lab: http://www.mouseatlas.org/data/mouse/libraries/SM149). While tags with single counts could be derived from bona-fide mRNA transcripts, they could also be generated through PCR or sequencing errors (Audic and Claveri, 1997; Siddiqui et al., 2005).  18  Thus, while the expression of hepatic transcription factors were detected in whole fetal liver SAGE libraries, the levels are below the threshold for meaningful analysis. Overall, transcriptome studies using whole fetal livers are inadequate for analyzing gene expression changes in differentiating hepatoblasts. To capture those gene expression changes, a transcriptome analysis that focuses specifically on hepatoblasts is required. As the expression of cell surface markers such as E-cadherin, MET, and DLK1 have been used to separate hepatoblasts from other cell types in the fetal liver (Nierhoff et al., 2005; Suzuki et al., 2002; Tanimizu et al., 2003), it is possible to first isolate hepatoblasts using a cell sorting strategy for the construction of gene expression libraries. However, to date, no such study has been reported.  1.7  Sequencing-based transcriptome analysis techniques: SAGE and Tag-seq Serial analysis of gene expression, or SAGE, is a powerful genome-wide  transcriptome analysis technique that employs sequencing technology (Hu and Polyak, 2006; Velculescu et al., 1995; Saha et al., 2002). There are two main principles behind the SAGE method. First, short sequence tags of 10-27 bp in length are generated at defined positions of the cDNA. This is achieved by capturing total mRNA via oligo(dT) beads, followed by cDNA synthesis, and subsequent cleavage with an anchoring enzyme. A common anchoring enzyme is NlaIII, a restriction enzyme that recognizes CATG sites that on average occurs every 256 bp in a genome (Pleasance et al., 2003). The NlaIII-digested fragments are ligated to adapters containing a CATG overhang and a recognition site for the tagging enzyme, which is a type IIS restriction enzyme that cuts a set distance away from the non-palindromic recognition site. Different restriction enzymes such as BsmFI, MmeI, and EcoP15I have been  19  used in different applications of the SAGE protocol (Harbers and Carninci, 2005), and cleavage of the cDNA-oligo(dT) fragments with this enzyme releases the short sequence tags (or SAGE tags) that will be used for downstream processing. The other main principle of SAGE is the concatenation, quantification, and identification of SAGE tags. After the tagging enzyme digestion step, SAGE tags are ligated to form ditags, which are PCR-amplified to generate sufficient material for downstream sequencing. The amplicons are subjected to an NlaIII digestion to remove adapter sequences, and the SAGE tags – flanked by the sequence CATG – are ligated to form concatemers before being cloned into a sequencing vector for capillary sequencing. The sequenced SAGE tags are mapped to curated transcript databases to determine their mRNA transcript of origin, and the number of times a particular SAGE tag is sequenced provides an absolute number for quantifying transcript expression level. In theory, each occurrence of a CATG (or the recognition site of the anchoring enzyme) within a cDNA can result in a SAGE tag. However, as SAGE captures mRNA via their polyA tails, SAGE libraries are dominated by tags nearest the 3’ end of mRNA transcripts (Velculescu et al., 1995). SAGE possesses several advantages over hybridization-based transcriptome techniques such as microarrays. Being a sequencing-based protocol, SAGE is not limited by the number of hybridization probes and can thus provide genuine genome-wide coverage (Velculescu et al., 1995). Also, SAGE requires no a priori knowledge of transcripts, which allows for the discovery of antisense and novel protein-coding transcripts (Chen et al., 2002; Ge et al., 2006; Nesbitt et al. 2010; Robinson et al., 2004; Wahl et al., 2005). Furthermore, the quantification of SAGE tags is a digital measure of transcript abundance. This feature allows SAGE data from different experiments to be cross-compared (Yang et al., 2010; Yu et  20  al., 2010), which greatly enhances the utility of the generated transcriptome libraries. The modular nature of the SAGE protocol also allows for easy modification of the protocol to suit specific needs. This is well demonstrated by the different versions of the technology since its inception: longSAGE (Saha et al., 2002), 5’ SAGE (Wei et al., 2004), microSAGE (Datson et al., 1999), and SuperSAGE (Matsumura et al., 2003). Lastly, SAGE possesses superior sensitivity over microarrays and is therefore better suited for detecting lowly expressed genes such as transcription factors (Lee et al., 2005; Hu and Polyak, 2006; Sun et al., 2004). The Hoodless lab has been extensively involved in the construction and analysis of longSAGE libraries. Specifically, in the Mouse Atlas of Gene Expression project (www.mouseatlas.org), over 200 SAGE libraries were constructed using embryonic, postnatal, and adult mouse tissues (Siddiqui et al., 2005). This project was a robust proof-ofconcept study showing SAGE to be an efficient method for profiling the transcriptome of different mouse tissues. To begin using the SAGE libraries to better understand embryonic development, libraries constructed from specific embryonic tissues were subjected to further analyses. First, libraries from the early mouse endoderm were compared against other Mouse Atlas SAGE libraries, and several novel endoderm-specific markers were characterized (Hassan et al., 2010; Hou et al., 2007). Using a similar approach, transcripts with enriched expression in the developing pancreas were found, which led to identification of regulatory cascades that underlie beta-cell development (Hoffman et al., 2008). Lastly, libraries from the developing heart were analyzed. By identifying transcripts that are differentially expressed in different heart compartments and with distinct temporal expression patterns, genes regulating heart valve formation were found (Vrljicak et al., 2010). Taken together,  21  these studies illustrate the power of SAGE libraries and how they have helped elucidate mechanisms that drive embryonic development. However, SAGE has several disadvantages that prevent it from being the ideal choice for transcriptome analysis (Butte, 2002; Hu and Polyak, 2006; van Ruissen et al., 2005). The steps in a conventional SAGE protocol are technically challenging and labour-intensive. Therefore, the technique is often impractical for small-scaled experiments with only a few samples or experimental timepoints. Furthermore, as the mouse genome consists of at least 2.5 x 109 base pairs (Mouse Genome Sequencing Consortium, 2002) and a tag sequence is only 10 to 26-bp in length, a tag mapping to multiple genomic locations cannot be used to definitively determine its gene of origin. Also, as the cost of capillary sequencing is expensive, SAGE libraries are often sequenced to a depth that is insufficient for genome-wide coverage (Hoffman et al., 2008; Hou et al., 2007; Morrissy et al., 2009; Vrlijicak et al., 2010; Wu et al., 2010). Lastly, the SAGE protocol can introduce technical artifacts arising from the PCR and sequencing steps, as well as a GC-bias in the SAGE tags (Morrissy et al., 2009). Recently, the SAGE protocol has been modified to couple with next-generation sequencing (NGS), resulting in a new technique called DGE (digital gene expression) or Tagseq (Figure 1-4). This technique is conceptually similar to SAGE, where mRNA is processed by restriction enzymes to generate 21-bp tags. But, instead of tag concatenation and cloning as in SAGE, the tags are first ligated to sequencing primers and then stabilized onto glass slides. After PCR amplification, the tags are then analyzed using NGS technology, allowing tags to be generated in a massively parallel fashion. Since its creation, many Tag-seq libraries have been generated using normal and malignant tissues (Morrissy et al., 2009; Wu et al., 2010). The Tag-seq technique solves several of the downfalls present in SAGE. Due to the 22  reduced cost of NGS (von Bubnoff, 2008), Tag-seq libraries can contain up to 10 million tags per library, which is a depth that allows for true genome-wide coverage of the transcriptome (Morrissy et al., 2009; Wu et al., 2010). The increase in sequencing depth also allowed the identification of developmentally regulated genes that escaped detection by LongSAGE (Ruzanov and Riddle, 2010). In addition, as tags are no longer concatenated and ligated into vectors, Tag-seq libraries are no longer biased toward GC-rich sequences (Morrissy et al., 2009).  1.8  Aim of study In this thesis work, I sought to perform a genome-wide, temporal analysis of the  hepatoblast transcriptome during mouse embryonic development. In Chapter 3, I describe the temporal and spatial expression patterns of the cell surface molecule, DLK1, in the developing fetal liver. The findings suggest DLK1 to be a specific marker for hepatoblasts and not other cell types in the fetal liver. Methods to facilitate the isolation of fetal liver DLK1+ cells for Tag-seq library construction are discussed. In Chapter 4, I present data from the Tag-seq libraries generated from fetal liver DLK1+ cells isolated at specific developmental timepoints. Analysis of the Tag-seq data showed high enrichment for hepatic genes, suggesting the hepatoblast transcriptome to be successfully captured by the Tag-seq libraries. Various TFs known to regulate hepatoblast differentiation were also detected, which is a significant improvement over existing fetal liver microarray and SAGE transcriptome libraries. Further analysis of the Tag-seq libraries revealed distinctive temporal gene expression signatures that underlie in vivo hepatoblast maturation. Gene Ontology analysis of the expression profiles showed a statistically significant enrichment of TFs in  23  E10.5 hepatoblasts. In Chapter 5, experiments focusing on the high mobility group A2, Hmga2, and its role in hepatoblast differentiation will be presented. Expression of the HMGA2 protein was abundant in nascent hepatoblasts but was drastically repressed in more mature hepatoblasts. Analysis in Hmga2-/- mutant embryos and liver cell lines indicate Hmga2 to potentially regulate both hepatoblast proliferation and differentiation. In Chapter 6, I discuss the overall significance of the findings in this thesis, as well as their application to other aspects of liver biology.  24  Cell type  Proportion of liver  Location in liver  Hepatocyte  70%  parenchyma  Cholangiocyte/ bile duct cell  3%  duct epithelium  Endothelial cell  10%  vasculature  forms blood vessels for proper liver zonation  Liver sinusoid endothelial cells  2.50%  sinusoids  allows metabolites transfer between hepatocytes and blood  Pit cell/liver natural killer cells  <1%  sinusoids  Kupffer cells/ liver macrophages  2%  sinusoids  Hepatic stellate cells  1%  perisinusoidal  Oval cells  <1%  duct epithelium  Functions glucose, protein, lipid, and urea metabolism; bile secretion transports bile to gall bladder  cytotoxic activity secrete cytokines and proteases extracellular matrix deposition; activated during regeneration bipotential progenitor cells; activated upon regeneration after severe liver injury  Table 1-1 Cell types in the adult liver. A normal adult liver also contains blood cells, which make up approximately 10% of the organ’s cell mass.  25  Figure 1-1 Architecture of the adult liver. Left panel: the mammalian liver is a highly structured organ that is organized in stereotypic hexagonal units. These units consist of hepatocyte cords radiating from the central vein and are infiltrated with hepatic veins to form the hepatic sinusoids. Right panel: at the corners of each hexagonal unit are the portal triad, made up of a hepatic portal vein, hepatic artery, and bile duct. Bile ducts function to transport bile synthesized by hepatocytes for storage in the gall bladder. Figure reproduced with permission from Gerard J. Tortora and Bryan Derrickson, “Introduction for the Human Body - 9th edition” (page 525), Wiley.  26  Figure 1-2 Early phases of hepatogenesis in the mouse embryo. Top panel: schematic of the first steps in mouse hepatogenesis. The process starts in the E8 embryo, when a portion of the ventral endoderm receives inductive signals from the adjacent septum transversum mesenchyme and heart mesoderm to commence expression of early liver markers by E8.5. At this stage, the mesenchymal tissues and endothelial cells also induce the outgrowth of hepatic progenitors. (VE = ventral endoderm; S = septum transversum mesenchyme; He = heart mesoderm; E = endothelial cells; Lb = liver bud; H = hepatoblast) Bottom panel: in situ hybridization of Hnf4! (A) and Gata4 (B) on transverse sections of an E9.5 embryo. Black arrow denotes location of the liver bud, while white arrows denote the septum transversum mesenchyme. Figure adapted from Roong Zhao and Stephen A. Duncan, “Embryonic development of the liver”, Hepatology 41 (5), pp. 956-967, May 2005. DOI: 10.1002/hep.20691 (http://onlinelibrary.wiley.com/doi/10.1002/hep.20691/abstract)  27  Figure 1-3 Bipotential hepatoblast differentiate into hepatocytes and cholangiocytes. Around E13.5, hepatoblasts begin to differentiate down either the hepatocyte or cholangiocyte lineage. This differentiation is accompanied by changes in gene expression as well as cellular morphology. Hepatocytes acquire apical-basal polarity and form tighter cellcell junctions with neighbouring hepatocytes. Cholangiocytes first organize themselves into a monolayer around the periportal vein, and sections of the monolayer are pinched off to form primitive bile ducts. These primitive hepatocytes and cholangiocytes continue to mature in the first week after birth. PV = periportal vein.  28  Figure 1-4 Tag-seq methodology. This technique couples the SAGE protocol with next-generation sequencing to generate deep sequencing transcriptome libraries. PolyA mRNA species from a cell population is isolated by polyT beads. After cDNA synthesis, the products are digested with NlaIII to generate fragments with GTAC overhangs. Adapter A containing a built-in MmeI site is ligated onto the overhangs, and digestion with MmeI cuts 21 base pairs downstream to remove the polyA tract and generate di-nucleotide overhangs. The Adapter A-tag sequence fragments are then gel purified, and Adapter B is ligated to the di-nucleotide overhangs to flank the tag sequence with distinct adapters. These adapters-tag sequence complexes are amplified by PCR before being read on a next-generation sequencer. Figure adapted with permission from Rebecca Cullum, Olivia Alder, and Pamela Hoodless, “The next generation: using new sequencing  29  technologies to analyse gene regulation”, Respirology 16 (2), pp. 210-222, February 2011, Wiley. DOI: 10.1111/j.1440-1843.2010.01899.x. (http://onlinelibrary.wiley.com/doi/10.1111/j.14401843.2010.01899.x/abstract;jsessionid=1307E0FDAA92D0172046BA4B3A6C8C9D.d03t02)  30  Chapter 2: Materials and methods  2.1  Mouse strains Mice were bred and maintained in the Animal Resource Centre of the BCCA Cancer  Research Centre, and all conditions met guidelines set by the Canadian Council on Animal Care. Inbred C57/BL6-J mice were used in the construction of Tag-seq libraries, while outbred ICR mice were used for immunohistochemistry, qRT-PCR, and primary culture experiments. B6.Cg-Hmga2pg-Tg40BCha/BmJ knockout mice (Benson and Chada, 1994; Xiang et al., 1990) were generously provided by Michael Copley and Dr. Connie Eaves (BCCA Cancer Research Centre). Embryos were obtained from mouse crosses using 6-8 week old females, and noon of the day with a vaginal plug was designated as embryonic day (E) 0.5.  2.2  Genotyping Hmga2 mutants Genomic DNA from adult ear punches or embryo limbs was extracted using ZyGEM  Gold and according to manufacturer’s protocol. 2!L of the genomic DNA was used as template for PCR genotyping reaction. The PCR conditions are: one cycle of 94°C for 3 minutes; 35 cycles of 94°C for 30 seconds, 60°C for 1 minute, and 72°C for 1 minute; and one cycle of 72°C for 2 minutes. See Table 2-1 for primer sequences.  2.3  Immunofluorescence microscopy Mouse embryos were fixed at 4oC overnight in 4% PFA (except for E16.5 embryos  which were fixed for two overnights), equilibrated through a 15-30-60% sucrose gradient (diluted with PBS), and embedded in a cryostat cassette with OCT medium (Tissue Tek).  31  4!m sections were obtained using the Leica CM3050S cryostat and post-fixed with 4% PFA (Sigma). Slides were washed twice in PBS containing 0.1% BSA (Roche) and 0.1% Tween20 (Sigma) at room temperature for 10 minutes, and blocked with PBS containing 5% BSA, 0.1% Tween-20, and 10% goat serum (Gibco) at room temperature for 90 minutes. Primary antibodies were added at the indicated concentrations (Table 2-2) in PBS, BSA, Tween-20, and goat serum and stained at room temperature for 60 minutes. Slides were washed 3-5 times at room temperature with PBS, BSA, and Tween-20 for 10 minutes each to remove excess antibody before the corresponding secondary antibodies were applied at room temperature for 60 minutes (Table 2-2). Slides were washed 3-5 times at room temperature for 10 minutes each, and DAPI (4',6-diamidino-2-phenylindole; Sigma) was added to the last wash at 1ug/mL to visualize nuclei. Slides were mounted in 50!L of 80% glycerol and 20mg/mL DABCO (1,4-diazabicyclo[2.2.2]octane; Sigma) and overlaid with a coverslip for viewing. Images were acquired using a Zeiss Axioplan fluorescent compound microscope and OpenLab software (PerkinElmer). For deconvolution, Z-stacks were obtained by acquiring 30-40 XY planes at 0.2!m intervals. Stacks were imported into the Volocity software (PerkinElmer) to perform iterative deconvolution using the 3D reconstruction module. For cytospins, dissociated fetal liver cells were affixed onto glass slides by spinning at 350 RPM for 5 minutes using the Shandon Cytospin 2. Cells were left to dry overnight and fixed with 4% PFA before proceeding with the antibody staining procedure described above.  32  2.4  Cell counting in immunofluorescent images To determine the degree of overlap between DLK1 and other cell types in the fetal  liver, sections co-stained with antibodies against DLK1 and markers of other cell types were imaged at 400x magnification. After image acquisition, two to three images containing approximately 1000 cells each were visually surveyed for two classes of cells: 1) co-stained against DLK1 and the cell type marker; and 2) stained against DLK1 alone. The number of double-positive cells were tallied and divided over the total number of DLK1-positive cells to generate the corresponding percentages. To compare HMGA2 and HNF4A staining intensity in HepG2 cells, the cells were grown on a circled area of a glass slide marked with wax pen. Cells were fixed with 4% PFA as described above, co-stained with antibodies against HNFA and HMGA2, and images at different fields of view were acquired at 400x magnification. Two images containing approximately 100 cells each were visually surveyed for twenty cells with the brightest staining against HNF4A, and twenty other cells with the brightest staining against HMGA2. From each category, eight cells were randomly selected for visual side-by-side comparison of the HNF4A and HMGA2 staining intensities.  2.5  Fetal liver dissociation Embryos of the desired stage were removed the uterine horns and dissected in cold  PBS. Livers were manually dissected from the embryo using a light dissecting microscope and fine forceps. Procedures for mechanical or enzymatic dissociation of mouse fetal livers have been previously described (Bowie et al., 2006; Miyazaki et al., 1981; Germain et al., 1988). Briefly, for mechanical dissociation, 6 E12.5 or 3 E14.5 fetal livers were placed onto a  33  40!m cell strainer (BD Biosciences), gently mashed up using a syringe plunger, and washed thoroughly with 1-2mL PBS into a collection tube. For enzymatic digestion, 8 E10.5 or 2 E16.5 fetal livers were placed in 2mL PBS containing 2% FBS (StemCell Technologies), 1!g/mL DNase (Sigma), 2.5 x 10-4 % collagenase (StemCell Technologies) and 2.4 mU/mL dispase (StemCell Technologies), and triturated at 37oC for 60 minutes. Liver clumps were dissociated by repeated pipetting, and the cell suspension was further dissociated by passing it through a 21.5 gauge needle. Cells were spun at 1100 RPM for 3 minutes at 4oC and washed once with PBS and FBS to remove excess enzymes before proceeding to next steps.  2.6  Flow cytometry analysis and FACS Flow cytometry analysis was done on the BD FACSCalibur, and fluorescence  activated cell sorting (FACS) was performed on the BD Influx and BD FACSAria cell sorters in the Flow Cytometry Core of the Terry Fox Laboratory at the BCCA Cancer Research Centre. For analysis or sorting of DLK1+ fetal liver cells, dissociated fetal liver cells were resuspended in PBS containing 2% FBS, 40!M EDTA and 1!g/mL DNase, and cells were stained with either the unconjugated or FITC-conjugated anti-DLK1 antibody (MBL International) for 10 minutes at 4oC. Cells were washed twice with 10-fold excess staining media and pelleted by centrifugation. If unconjugated anti-DLK1 antibody was used, cells were stained with either anti-rat PE (Rockland) or anti-rat Alexa Fluor 488 (Invitrogen) secondary antibody for 10 minutes at 4oC. 1!g of PI (propidium iodide; Sigma) was added to the last wash to discriminate dead cells during FACS, and the resuspended cells were passed through a 40!m cell strainer before FACS analysis or sorting. Dissociated fetal liver cells that were either unstained or stained with only the secondary antibody were used as negative  34  controls. For sorting, only live and singlet cells were gated. Sort purity was assayed by FACS re-analysis of DLK1+ cells, and cell populations of at least 90% purity were used for library construction. To pool sorted cells for library construction, cells were pelleted by spinning at 4000RPM for 5 minutes at 4oC, and stored at -80oC in Trizol (Invitrogen). For sorting HPPL cells after lentiviral infection, untransfected HPPL cells were used as negative controls, and live, singlet cells positive for MIY or Hmga2-MIY were sorted directly into Trizol for RNA extraction.  2.7  Magnetic activated cell sorting For magnetic isolation of DLK1+ cells, EasySep mouse FITC- or PE selection kits  from Stem Cell Technologies were used accordingly to the manufacturer’s protocol, with the following modifications. Dissociated fetal liver cells resuspended in 1.2mL of PBS containing 2% FBS, 40!M EDTA, and 1!g/mL DNase were stained with either FITC- or PE-conjugated anti-DLK1 antibody for 10 minutes at 4oC. Cells were washed with 10-fold excess of the PBS-based staining medium, pelleted by spinning at 1200 RPM for 3 minutes at 4oC, and resuspended in 1.5mL of staining medium. Cells were then stained with 110!L of the FITC or PE selection cocktail for 15 minutes at 4oC, followed by incubation with 80!L nanomagnetic particle for 10 minutes at 4oC. This final tube containing magnetic DLK1+ cells was brought to a final volume of 2.5mL before being placed into the EasySep magnet for 5 minutes. Non-magnetic cells were removed by decanting, and cells retained by the magnet were rinsed with another 2.5mL of staining medium and placed in the magnet for an additional 5 minutes.  35  2.8  Tag-seq library construction Approximately 1!g of RNA was collected for construction of each Tag-seq library  (Table 4-1). Tissue from the adult liver was placed in 1mL Trizol and pulverized with a hand homogenizer. If pooling of FACS-isolated cells was required, cells were pelleted and stored in 200!L of Trizol at -80oC, and when sufficient material had been collected, the pooled cells were placed in a final volume of 1mL Trizol and delivered to the Michael Smith’s Genome Sciences Centre at the BCCA Cancer Research Centre for construction of Tag-seq libraries (Figure 1-4). First, mRNA was captured by oligo(dT) beads, followed by cDNA synthesis and digestion with the NlaIII anchoring enzyme. This generates a 4-bp overhang that is ligated to Adapter A, which contains an MmeI restriction enzyme site that cuts 21 bp downstream from the recognition site. After an MmeI digest, the cDNA fragments were purified and ligated to Adapter B. These fragments were PCR amplified using primers against the flanking adapters, and the amplified products were sequenced on the Illumina Genome Analyzer to generate 21-bp tags. Data from the Tag-seq libraries are stored in the Genome Sciences Centre data depository and are available upon request.  2.9  Mapping tags and normalizing libraries To exclude tag sequences that were either lowly expressed or generated through PCR  and other technical errors, the SSOOHE (sans singletons and one-offs of highly expressed tags; Morrissy, 2010) filtering algorithm was applied to Tag-seq libraries before mapping. This filtering algorithm removes tag sequences that are observed only once (ie. singletons), or if their sequence contains a 1-bp mismatch from a second, more highly expressed tag (ie. one-offs of highly expressed tags). SSOOHE tags were mapped to transcripts in the mouse  36  RefSeq mm8 assembly (Pruitt et al., 2007) using the DiscoverySpace 4 program (Robertson et al., 2007). Tags mapping to the sense or antisense transcript were denoted as having positive or negative positions, respectively, with tags at the +1 or -1 positions being closest to the transcript 3’ end. A tag was considered to be uniquely mapping if it is found in only one entry of the database. Raw tag counts were normalized by dividing over the cumulative counts of all SSOOHE tags in that library, and then multiplied by 1,000,000 to obtain normalized tag counts expressed in tags per million (TPM).  2.10 K-means clustering In this study, we focused on all positive positions tags that map uniquely to the RefSeq mm8 database. To include all tags derived from a gene, tags mapping to different positive positions of a gene were collapsed and their normalized tag counts summed. A gene was retained for clustering analysis if it was present with at least 10 TPM in one of the Tagseq libraries. Applying this threshold yielded a total of 5449 genes. To convert normalized expression levels to values suitable for K-means clustering, the Tag-seq data was manipulated a second time to generate relative expression levels. To achieve this, normalized tag counts for each gene was summed across all timepoints, and the counts at a given timepoint was divided by that sum to generate a number between 0 and 1. To determine the optimal cluster number for K-means analysis, the relative expression level of all 5449 genes were subjected to a figure-of-merit test using the MeV program (Saeed et al., 2003). The figure-of-merit graph showed a plateau when K is equal to 12, and this value was chosen for subsequent K-means clustering analysis.  37  Using the K-means clustering algorithm (Cai et al., 2004), the 5449 genes were grouped into 12 clusters based on their temporal expression patterns in the Tag-seq libraries, and consensus clusters after 100 iterations were visually inspected. Seven clusters with clean temporal patterns were retained as is, two clusters with highly similar patterns were merged into one, and three clusters with noisy patterns were merged together and subjected to a second round of clustering. Re-clustering with 50 iterations yielded six additional clusters, generating a total of 14 gene clusters overall.  2.11 Gene Ontology analysis The DAVID Functional Annotation tool (http://david.abcc.ncifcrf.gov/summary.jsp; Huang et al., 2007) was used to identify Gene Ontology categories enriched in a given gene cluster. Genes from each of the 14 gene clusters were used as the input target gene set, while all 5449 genes subjected to the K-means clustering analysis were used as background. Gene Ontology categories were considered to be enriched if the fold enrichment is above 1.45 with a Bonferroni value below 0.05.  2.12 cDNA synthesis and qRT-PCR cDNA synthesis and qRT-PCR were performed as previously described (McKnight et al., 2007). Approximately 1!g total RNA was treated with DNAase I, and 0.1!g of the DNAase I-treated RNA was added with 6!g of random hexamers and 200U Superscript II Reverse Transcriptase (Roche) to synthesize cDNA in a volume of 20!L. The resulting cDNA solution was diluted to a final volume of 50!L, from which 1.5!L was used per qRTPCR reaction. qRT-PCR was performed using the Roche Universal SYBR Green mix and the  38  ABI 7900HT Fast Real-time PCR System. The cycling conditions were 95oC for 10 minutes, followed by 40 cycles of 95oC for 30 seconds, 55oC for 30 seconds, 72oC for 30 seconds, and a final extension time of 72oC for 10 minutes. Technical triplicates were performed for each biological sample; see Table 2-1 for qRT-PCR primer sequences.  2.13 Lentiviral infection The Hmga2 lentivirus was a generous gift from Michael Copley and Dr. Connie Eaves. The Hmga2 cDNA was cloned into the pCCL.PPT.MND.PGK.YFP lentiviral vector (Challita et al., 1995). The vector was transfected into HEK293T cells using a standard calcium phosphate protocol, and media containing lentiviral particles were collected, filtered, and concentrated as previously described (Imren et al., 2004). To infect HPPL cells, Hmga2 lentivirus was added to the media at a concentration of 1:100 and the empty vector control lentivirus at a concentration of 1:50. After three hours, the infected cells were washed three times with 10-fold excess DMEM before returning cells to their regular growth media.  2.14 Cell culture All cell cultures were maintained in a 37oC incubator with 5% CO2. HepG2 cells were maintained in DMEM (StemCell Technologies) and 10% FBS. HPPL cells were derived from E14.5 DLK1+ cells as previously described (Tanimizu et al., 2003). First, E14.5 DLK1+ cells were isolated by MACS, plated at 2000 cells/cm2 on laminin-coated 6-well plates, and cultured in a DMEM/F12 solution containing 2 mM L-glutamine (Gibco), 10% FBS, 1x nonessential amino acid (StemCell Technologies), 1x ITS (Gibco), 10–7 M Dex (Sigma), 20ng/mL EGF (Peprotech), and 20ng/mL HGF (Peprotech). Media was changed  39  every 4-5 days and cells were kept in culture for two months, by which time the cells can be lifted, replated, and passaged as a cell line.  40  Gene HMGA2tg * HMGA2KO ** oIMR8744/5 # Afp  "-actin Bambi Egr1 Hmga2 Hnf4a Igf2r  Forward primer 5’-ATT CTG GAG ACG CAG GAA GA-3’ 5’- CCC ACT GCT CTG TTC CTT GC -3’ 5’- CAA ATG TTG CTT GTC TGG TG -3’ 5’-GAA GCA AGC CCT GTG AAC TC-3’ 5’-CCA GAG CAA GAG AGG TAT CCT G-3’ 5’-ATG CCC ACT TTG GAA TGC TGT CAC-3’ 5’-ACG AGT TAT CCC AGC CAA ACG ACT-3’ 5’-GGT GCC ACA GAA GCG AGG AC-3’ 5’-GAT TGC CAA CAT CAC AGA CG-3’ 5’-GGT GAA GCC CAA TGA CCA GCA TTT-3’  Reverse primer 5’-TGC TCC TGG GAG TAG ATT GG -3’ 5’-GTG TCC CTT GAA ATG TTA GGC G -3’ 5’-GTC AGT CGA GTG CAC AGT TT -3’ 5’-GGC ATA GGT TTC ATC CCT CA-3’ 5’-CAT TGT AGA AGG TGT GGT GCC AG-3’ 5’-TGC TGA TAC CGG TTT CCT TGT CCT-3’ 5’- ACA AAG TGT TGG CAC TGT TGG GTG-3’ 5’-GGG CTC ACA GGT TGG CTC TT-3’ 5’-CAG AAG GAG TTC GCA GAA GG-3’ 5’-TTT CCC ACA CTT GGA GGA AGA GGT-3’  Table 2-1 Primers used in genotyping and qRT-PCR. * = genotyping primers against artificial locus present in the animals carrying the Hmga2 KO cassette; primer pair generates a 369-bp amplicon ** = genotyping primers against the first exon of Hmga2; primer pair generates a 288-bp amplicon # = genotyping internal control against a locus on chromosome 14; primer pair generates a 200-bp amplicon  41  Primary antibodies  Antibody against  Host species  DLK1  rat  DLK1FITC  rat  DLK1-PE  rat  DLK1 HNF4A HNF4ADyLight 488 CD45 Ter119FITC * CD45-PE *  Secondary antibodies  Company  Catalogue #  Final concentration  D187-3  1/400  D187-4  1/200  D187-5  1/200  rabbit rabbit  MBL International MBL International MBL International Chemicon Santa Cruz  AB3511 sc-8987  1/600 1/100  rabbit  Epitomics  4580-1  1/50  rat  BD Biosciences  561088  1/100  rat  eBioscience  115921  1/100  rat  eBioscience BD Biosciences Lab Vision Cell Signaling  140451  1/100  550274  1/150  MS-376-30  1/150  5269  1/250  PECAM  rat  Desmin  mouse  HMGA2  rabbit  rat PE  rabbit  Rockland  RL7124826  1/1000  goat  Invitrogen  A11006  1/1000  donkey  Invitrogen  A21207  1/1000  goat  Invitrogen  A11017  1/1000  rat Alexa Fluor 488 rabbit Alexa Fluor 594 mouse Alexa Fluor 488  Table 2-2 Antibodies used for immunofluorescence microscopy and FACS. * = antibodies used in preliminary studies for data not shown in this thesis  42  Chapter 3: Characterization of DLK1+ cells in the mouse fetal liver  3.1  DLK1+ cells are present in the fetal liver The overarching aim of this thesis was to perform a temporal transcriptome analysis  during in vivo hepatoblast maturation. However, as the fetal liver is predominantly hematopoietic (Crawford et al., 2010; Sigal et al., 1994), hepatoblasts first need to be isolated from the fetal liver to facilitate such analysis. Various cell surface markers specific to mouse fetal hepatoblasts have been identified, such as E-cadherin, MET, and DLK1 (Nierhoff et al., 2005; Suzuki et al., 2002; Tanimizu et al., 2003). After the initial publication that identified DLK1 as a cell surface marker on E14.5 hepatoblasts, a number of studies using primary cultures and the HPPL cell line derivative have confirmed DLK1+ fetal liver cells to be bipotential progenitors (Tanimizu and Miyajima, 2004; Tanimizu et al., 2004; Tanimizu et al., 2007; Tanaka et al., 2009). Therefore, as DLK1+ hepatoblasts have been more thoroughly characterized than hepatoblasts expressing other cell surface molecules, DLK1 was selected as the hepatoblast marker of choice. First, to confirm DLK1 is expressed during fetal liver development, E10.5 to E16.5 fetal livers were homogenized, fixed, and stained with an anti-DLK1 antibody. Flow cytometry analysis confirmed that DLK1+ cells were present during liver development, but at varying percentages (Figure 3-1). Close to a fifth of E10.5 fetal liver cells expressed DLK1, and this figure drops to just below 15% by E12.5. The proportion of DLK1+ cells continues to drop at E14.5 and E16.5, decreasing by approximately half every two days. While the proportion of DLK1+ cells showed a progressive decrease during the course of development,  43  these data nonetheless suggest DLK1 to be continuously expressed in the fetal liver after midgestation.  3.1.1  DLK1 is expressed on hepatic cells in the fetal liver Next, to determine if the DLK1 cell surface protein is specific to hepatoblasts, the  expression pattern of DLK1 was examined against that of a well-characterized hepatoblast marker. The HEX and HNF4A transcription factors as well as the AFP and ALB serum proteins have been routinely used to label hepatoblasts in the fetal liver (Bort et al., 2006; Germain et al., 1988; Zhao and Duncan, 2005). Among these, the function of HNF4A has been the most extensively characterized in the context of fetal liver development. Hnf4! mRNA is first detected in nascent hepatoblasts of the E9.0 embryo, and the expression persists over the course of fetal liver development (Taraviras et al., 1994). Hnf4! regulates hepatoblast differentiation by activating numerous hepatic TFs and cell adhesion molecules needed for proper hepatocyte differentiation (Battle et al., 2006; Li et al., 2000; Parviz et al., 2003). In addition, as Hnf4! is a transcriptional target of other TFs that control hepatoblast differentiation, the absence or presence of Hnf4! expression has been used in numerous mutant mice models to indicate a loss or gain of hepatocyte differentiation (Hunter et al., 2007; Krupczak-Hollis et al., 2004; Ludtke et al., 2009; Qu et al., 2007). For these reasons, HNF4A was chosen as the hepatoblast marker for this study. To examine whether fetal liver cells expressing DLK1 also express HNF4A, sagittal sections from embryos of different stages were co-stained with antibodies against DLK1 and HNF4A. DLK1 expression during hepatogenesis was first detected in the E9.5 liver bud. DLK1+ cells were stretched and polygonal (Figure 3-2), which is characteristic of the  44  migrating hepatoblasts at this stage (Shiojiri, 1981; Wilson et al., 1963). More importantly, DLK1 was already localized to nascent hepatoblasts, as 75.9% (85/113) of DLK+ cells stained positive for nuclear HNF4A (Table 3-1). To extend this finding to later timepoints, the same co-staining experiments were performed on E10.5, E12.5, E14.5, and E16.5 fetal liver sections. As shown in Figure 3-3, the majority of DLK1-positive cells also expressed HNF4A. At E10.5, DLK1+ cells were connected and formed clusters reminiscent of proliferating hepatoblasts (Shiojiri, 1981; see Figure 3-3A), and 89.8% (325/362) of E10.5 DLK1+ cells expressed HNF4A (Table 3-1). In E12.5 and E14.5 fetal livers, cells expressing DLK1 became interspersed but still resembled irregular polygons (Figure 3-3D and G), and the proportion of DLK1+ cells expressing HNF4A remained steady at 86.2% (304/353) and 90.2% (370/410), respectively. At E16.5, DLK1+ cells were more rounded in morphology (Figure 3-3J). In addition, only 74.0% (213/288) of DLK1+ cells had HNF4A staining (Table 3-1). This finding could be explained by one of two possibilities. In the first scenario, all hepatic cells express the DLK1 antigen, and roughly a quarter of E16.5 hepatic cells have lost HNF4A expression (Figure 3-3J, white arrows). Alternatively, all hepatic cells express the HNF4A transcription factor, and roughly a quarter of E16.5 DLK1+ cells are non-hepatic. In the late-gestation embryo, Dlk1 is expressed in endodermal organs such as the liver and pancreas, as well as multiple tissues of mesoderm origin such as the lung, limb bud, and vertebrae (Carlsson et al., 1997; Floridon et al., 2000; Smas and Sul, 1993; Tanimizu et al., 2003; Wang et al., 2006). In contrast, Hnf4a expression in the post-midgestation embryos is mainly restricted to endoderm-derived tissues such as the liver, pancreas and colon, with kidney being the only mesoderm-derived organ expressing Hnf4a (Duncan et al., 1994; Garrison et al., 2006). More  45  importantly, much is known about the roles of HNF4A during hepatocyte differentiation (Battle et al., 2006; Li et al., 2000; Parviz et al., 2003). While DLK1 has been used as a hepatoblast marker, its role during the hepatocytes differentiation process is still unclear (Moon et al., 2002; Tanimizu et al., 2003). Therefore, expression of HNF4A is a more justified choice as a hepatic marker than DLK1, and thus implies DLK1+ HNF4A- cells to be non-hepatic. Overall, our antibody staining results indicate that the DLK1 cell surface molecule is present in the developing fetal liver. More importantly, DLK1 is an excellent marker for labeling E10.5, E12.5, and E14.5 hepatoblasts, as the percentage of fetal liver DLK1+ cells co-expressing HNF4A is close to 90% at those timepoints. In the E16.5 fetal liver, approximately a quarter of DLK1+ cells do not express HNF4A and are thus deemed nonhepatic. While the identity of these non-hepatic DLK1+ cells is yet unclear, the majority of E16.5 fetal liver DLK1+ cells are hepatic as close to 75% of the cells co-expressed HNF4A. Since DLK1 is an antigen that is predominantly expressed on hepatoblasts at the developmental stages observed, we decided to further pursue the utility of DLK1 as a cell surface molecule for labeling hepatoblasts during the fetal liver development.  3.1.2  Other fetal liver cell types do not express DLK1 After demonstrating DLK1 to be expressed on hepatic cells, we next asked if DLK1  expression was excluded from other cell types that are present in the fetal liver. First, DLK1 expression was compared against that of leukocyte common antigen/CD45, which is found on myeloid progenitors in the fetal liver but absent on differentiating erythroblasts (McCarthy et al., 1985; Uchida et al., 2001). In the fetal liver, hepatoblasts and myeloid  46  progenitors reciprocally support each other’s expansion. For instance, E14 CD45+ fetal liver cells secrete the oncostatin M cytokine to induce hepatocyte differentiation in vivo (Kamiya et al., 1999), while E15.5 SCF+ DLK1+ fetal liver cells secrete angiopoietin 3 and IGF2 to support expansion of hematopoietic stem cells (Chou and Lodish, 2010). While CD45+ cells are known to be present in the fetal liver, a description of the morphology of CD45+ cells in the native fetal liver environment as well as their spatial relationship with hepatoblasts is lacking. To examine this, E10.5 to E16.5 liver sections were co-stained with antibodies against DLK1 and CD45. As shown in Figure 3-4, the expression pattern of CD45 showed a punctate pattern, and the expression was localized to fetal liver cells that are small and rounded (Figure 3-4B, E, H, K). Also, CD45+ cells were found to be adjacent to DLK1+ hepatoblasts, which was most readily observed in the E12.5, E14.5, and E16.5 livers (Figure 3-4F, I, L). Significantly, the expression of CD45 did not overlap with DLK1 during any of the analyzed timepoints of fetal liver development. These data suggest that while DLK1+ and CD45+ cells are in close proximity in the fetal liver, the two populations are mutually exclusive. Next, we investigated the expression of DLK1 in the fetal liver relative to PECAM, a protein expressed on endothelial cells. Similar to fetal liver CD45+ myeloid cells, these cells are required for proper hepatogenesis. Specifically, they regulate an early step during hepatogenesis as mutants without endothelial cells have liver buds that fail to invade the septum transversum mesenchyme (Matsumoto et al., 2001). Endothelial cells have a stretched morphology and form tubular structures (Cherqui et al., 2006), which can be visualized in the fetal liver by labeling with the endothelial cell-specific markers LYVE1 or PECAM (Matsumoto et al., 2001; Takabe et al., 2012; see Figure 3-5B, E, H, and K). While  47  LYVE1 or PECAM-positive endothelial cells have been proposed to interact with hepatoblasts in the fetal liver, the exactly spatial relationship of this interaction has not been characterized. To achieve this, fetal liver sections were co-stained with anti-DLK1 and antiPECAM antibodies. At E10.5 and E12.5, PECAM+ cells were close but not immediately adjacent to DLK1+ cells (Figure 3-5C, F). However, in E14.5 fetal livers, PECAM and DLK1-positive cells were found as neighbouring populations (Figure 3-5I), but no overlap of expression was observed. A previous study has shown PECAM expression to be limited to hepatic veins and arteries in E17.5 fetal livers (Takabe et al., 2012), a similar pattern that was observed in E16.5 fetal livers (Figure 3-5K). Note that while some PECAM staining appeared to overlap with that of DLK1 (Figure 3-5L), the diffuse glow of PECAM staining in the overlapping regions does not resemble endothelial cells and thus is likely to be nonspecific. Overall, during fetal liver development, PECAM+ and DLK1+ cells are often observed in close proximity with each other, but the populations do not overlap. Lastly, we compared the expression of DLK1 and desmin in E10.5 to E16.5 fetal livers. Rat fetal liver cells that are adherent upon culturing express low levels of desmin and alpha smooth muscle actin, suggesting some hepatocytes to express markers for mesenchyme cells (Mansuroglu et al., 2009). Desmin encodes a central subunit of the intermediate filaments in muscle cells (Paulin and Li, 2004). This protein is used to label sickle-shaped stellate cells in the fetal liver, and studies have suggested this cell type to physically interact with fetal hepatoblasts (Nitou et al., 2000; Takabe et al., 2012). To characterize the interaction of stellate cells with cells expressing DLK1, fetal liver sections were stained with anti-DLK1 and anti-desmin antibodies, and the staining was analyzed by fluorescent microscopy. Similar to previous studies, cells expressing desmin were small and sickle-  48  shaped (Figure 3-6B, F, J, and N). Interestingly, desmin-positive cells showed slight overlap with DLK+ cells throughout liver development (Figure 3-6C, G, K, and O). The greatest overlap was observed at the E10.5 and E12.5 timepoints, where 14.0% (24/171) and 10% (24/240) of fetal liver DLK1+ cells showed desmin expression, respectively (Table 3-1). The overlap is reduced to 2.3% (5/220) in the E14.5 and 5.7% (12/209) in E16.5 fetal livers. In summary, the array of immunofluorescent staining experiments presented here has conclusively shown the DLK1 cell surface protein to be predominantly expressed on HNF4A-positive hepatoblasts during fetal liver development (Table 3-1). More importantly, no overlap was observed between DLK1 and the endothelial marker PECAM and the myeloid progenitor cell marker CD45. However, DLK1 expression showed minor overlap with the stellate cell marker desmin. Subsequent Tag-seq libraries generated with fetal liver DLK1+ cells can be used to determine the transcript levels of desmin as well as other mesenchyme genes in developing DLK1+ cells.  3.2  Enzymatic dissociation of fetal livers is optimal for FACS isolation of DLK1+ cells Results from antibody staining experiments have demonstrated DLK1 to be an  appropriate marker for labeling fetal hepatoblasts. To achieve our goal of capturing the hepatoblast transcriptome, we sought to isolate fetal liver DLK1+ cells by FACS. This is the method of choice for obtaining high-grade RNA from cell populations at purity of 90-95% (Barrett et al., 2002; Davies, 2007), both of which are essential elements for generating highquality transcriptome libraries. Two methods of obtaining single cell suspensions from fetal livers have been described. The first method is mechanical dissociation, where livers are dissociated by manual pipetting and forcing the mixture through a 40 !m mesh (Wolber et  49  al., 2002), and the method is mostly applied for hematopoietic studies (Bowie et al., 2006). The second method is enzymatic dissociation, where whole livers are incubated at 37oC in a saline-buffered solution containing collagenase and dispase (Cantz et al., 2003; Germain et al., 1998). This method is more frequently used for extracting hepatic cells for further experiments (Tanimizu et al., 2003; Kamiya et al., 1999). To determine which method is most efficient for isolating DLK1+ cells, fetal livers were dissociated with each technique, stained with an anti-DLK1 antibody against an extracellular epitope of DLK1, and DLK1+ cells isolated using FACS. When E12.5 livers were dissociated mechanically, just over 0.72% of the initial live, single-cell suspension stained positive for DLK1 (Figure 3-7B). When these DLK1+ cells were sorted and the resultant population checked for purity, the proportion of DLK1+ cells increased to approximately 40% (Figure 3-7C). While this is a significant enrichment from the initial population, the purity is suboptimal for constructing transcriptome libraries. Therefore, the DLK1+ cells were re-sorted to obtain cells of >90% purity (Figure 3-7D). Next, E12.5 livers were triturated using the enzymatic dissociation method. This method yielded DLK1+ cells at close to 4.4% in the initial live cell population (Figure 3-7F). This number is in accordance with previous studies that showed E12.5 DLK1+ fetal liver to be present at 2-10% (Tanaka et al., 2009; Tanimizu et al., 2003). Note that this percentage is smaller than the 14.9% observed in fixed E12.5 fetal livers (Table 3-1), because the fixation process precludes the discrimination of dead or live cells in subsequent FACS analysis. More importantly, the yield of live DLK1+ cells obtained after enzymatic dissociation was almost seven times more than the yield obtained from mechanical dissociation, and when enzymatically dissociated liver cells marked by the anti-DLK1 antibody were sorted, DLK1+  50  cells of >90% purity were obtained after just one round of sorting (Figure 3-7G). These results indicate that enzymatic dissociation is the more efficient method for FACS isolation of fetal liver DLK1+ cells.  3.3 3.3.1  Discussion Expression of DLK1 is observed predominantly on hepatoblasts but is also  present on non-hepatic cells In this chapter, I characterized the spatial expression pattern of the DLK1 cell surface molecule relative to markers of other cell types commonly found in the fetal liver. First, I showed that the DLK1 cell surface protein is expressed on hepatoblasts throughout the course of fetal liver development (Table 3-1). Specifically, the percentage of DLK1+ cells coexpressing the hepatic-specific transcription factor HNF4A was close to 90% in the E10.5, E12.5, and E14.5 fetal livers. However, in E16.5 livers, the percentage of DLK1+ cells with HNF4A expression dropped to 74%. Bipotential hepatoblasts begin their differentiation into hepatocytes or cholangiocytes around E13.5 (Germain et al., 1988). At E14.5, hepatoblasts still express Hnf4a (Tanimizu et al., 2003), but the expression is lost in hepatoblasts that commit toward the cholangiocytic lineage (Yamasaki et al., 2006). This raises the possibility that E16.5 DLK1+ HNF4A- cells are cholangiocytes. However, this model contrasts with previous in situ hybridization experiments that showed no expression overlap between Dlk1 and the cholangiocyte marker CK19 (Tanimizu et al., 2003). Further experiments comparing the expression of Dlk1 against additional cholangiocyte markers such as gamma-glutamyltranspeptidasese (Vroman and LaRusso, 1996) can be performed to confirm the absence of Dlk1 expression on cholangiocytes.  51  DLK1 was originally identified as PREF-1, a transmembrane protein that is highly expressed in preadipocytes (Smas and Sul, 1993). In mice, adipose tissue begins to develop in the perinatal embryo, and the first preadipocytes are found beneath the dermis two to three days before birth (Wojciechowicz et al., 2008). However, the origin of these preadipocytes is still unclear (Gregoire et al., 1998). Interestingly, some traits are shared between preadipocytes and DLK1+ cells in the E16.5 fetal liver. For instance, they both have a rounded cellular morphology (Gregoire et al., 1998; Figure 3-3J). Also, CEBPs and PPARs are two classes of TFs that play roles in both hepatocyte and adipocyte differentiation (Li et al., 2004; Darlington et al., 1998; Panadero et al., 2000; Schrem et al., 2004), suggesting hepatocytes and preadipocytes to have a similar transcriptional repertoire. Therefore, it is possible that some E16.5 DLK1+ HNF4A- cells are preadipocytes. One method to investigate this further is to perform oil red O labeling in the fetal liver to determine if DLK1+ HNF4Acells have lipid-storing capabilities (Li et al., 2004). In addition, E16.5 DLK1+ cells can be isolated and assayed for sustained expression of preadipocyte chemokines such as CCL2, SERPINE1, and IL6 upon in vitro culture (Hauner, 2005; Mack et al., 2009; Trayhurn, 2005; Wellen and Hotamisligil, 2005). However, as DLK1+ HNF4A- cells constitute approximately 25% of the total DLK1 population in E16.5 fetal livers, only a subset of cells subjected to this experiment may be responsive to the in vitro assay. Conversely, it is possible for DLK1+ HNF4A+ cells to also secrete preadipocyte chemokines in culture, thereby complicating data interpretation. Identifying cell surface molecules that specifically mark HNF4A-negative DLK1+ cells would greatly enhance the resolving power of the described assay. At the timepoints analyzed in this study, DLK1 showed no overlap with CD45, a protein expressed specifically on myeloid progenitors in the fetal liver (Figure 3-4). This  52  finding is in disagreement with a study that detected DLK1+ CD45+ double positive cells in the E10.5 and E12.5 fetal livers at 6.7% and 1.9%, respectively (Khurana and Mukhopadhyay, 2008). The discrepancy is likely a result of dissimilar dissociation and cell sorting methods. Khurana and Mukhopadhyay dissociated fetal livers mechanically by forcing them through a 23-gauge needle, and then sorting for DLK1+ cells using magnetic activated cell sorting (MACS). While MACS is a gentler protocol than FACS in terms of cell handling, its sorting parameters are less rigorous (Davies, 2007). For instance, MACS does not allow for the exclusion of cellular debris, doublet, or dead cells, which are all factors that can increase contamination. Therefore, the observed DLK1+ CD45+ double positive cells could be an artifact due to a combination of insufficient dissociation of fetal liver cells and limitations of the MACS protocol. Our antibody staining results in fetal liver sections showed a small degree of overlap between DLK1 and desmin, a stellate cell marker (Table 3-1). This finding may suggest desmin to be expressed on a small subset of hepatoblasts. Alternatively, as a subset of DLK1+ cells does not express the HNF4A hepatic marker, it could be the same subset of DLK1+ cells that are desmin-positive. The expression pattern of desmin in the developing mouse fetal liver has been documented in detail. Using an anti-desmin antibody to stain E9.5 liver sections, desmin was detected in the septum transversum mesenchyme but was completely absent in the neighbouring, E-cadherin-positive hepatoblasts (Nitou et al., 2000). When E11.5 liver sections were analyzed, E-cadherin-positive hepatoblasts were still distinct from desmin-positive stellate cells (Nitou et al., 2000; Takabe et al., 2012). As evidence presented by Nitou et al. and Takabe et al. suggest hepatoblasts to lack desmin expression, DLK1+ cells that express desmin are likely to be non-hepatic. Further experiments can be  53  performed to better characterize the DLK/desmin co-expression pattern in the fetal liver. For instance, the DLK1/desmin co-staining experiments could be repeated and the specimen imaged by confocal microscopy. The superior resolution of confocal imaging should aid in discerning staining that was difficult to distinguish by conventional compound microscopes. In addition, embedding samples in paraffin wax, a medium that better preserves histology than the OCT compound (Spector and Goldman, 2006), could also help prevent tissue shearing that could complicate image acquisition.  3.3.2  The method of fetal liver dissociation dictates the number of FACS rounds  needed for isolating DLK1+ cells to a purity of 90% or higher Two methods of fetal liver dissociation (mechanical or enzymatic) were tested for their efficiency in isolating highly purified DLK1+ cell. As the isolated DLK1+ cells will be ultimately used for transcriptome analysis, care must be taken in minimizing RNA degradation during the isolation procedure. Several factors could affect RNA quality, such as temperature and length of time at which the target cells are handled and stored during processing (Kasahara et al., 2006). Thus, for both the mechanical and enzymatic methods of dissociation, fetal liver samples were kept at 4oC and processed as quickly as possible. However, inherent differences exist between the two methods that could affect RNA quality. For instance, for mechanical dissociation, bulk fetal liver tissue is converted to single-cell suspensions using chilled instruments and reagents, and the process is complete in under two minutes. In contrast, for enzymatic dissociation, fetal livers are incubated with digestive enzymes and phosphate buffered saline at 37oC for one hour. These conditions are considerably warmer and more time-consuming relative to mechanical dissociation and thus  54  may cause increased RNA degradation. Similarly, the difference in the time needed to isolate DLK1+ cells by FACS could also affect final RNA quality. High purity DLK1+ cells can be obtained after just one round of FACS using enzymatically dissociated fetal liver cells, whereas an additional round of sorting is required for fetal liver cells that are mechanically dissociated (Figure 3-7). Here, the enzymatic dissociation method is advantageous, as less total time is needed for the cell sorting step. Further experiments to rigorously examine RNA quality after both dissociation methods would help determine the optimal choice for isolating hepatoblasts and preserving their RNA for transcriptome analysis. Two additional discrepancies surrounding the detection and isolation of DLK1+ in the fetal liver are worth noting. First, for E12.5 fetal livers that were enzymatically dissociated and fixed for flow cytometry analysis, DLK1+ cells were detected at close to 15% of total cells. Conversely, when E12.5 fetal liver were similarly processed but not fixed in order to facilitate live cell sorting, DLK1+ cells were present at the lower percentage of 4.4%. As hepatoblasts mature, they form increasingly strong cell-cell junctions with surrounding cells (Iwai et al., 2000). In order to obtain single-cell suspensions, these connections need to be disrupted, which likely causes shearing and death to some hepatoblasts. Therefore, the percentage of DLK1+ cells observed during cell sorting experiments represent the hepatoblasts that are still alive after dissociation, and it is likely an underestimate of the total number of hepatoblasts present in the fetal liver. Secondly, the percentage of DLK1+ cells in the initial population of dissociated cells greatly differs between the two methods of fetal liver dissociation. DLK1+ cells make up less than 0.72% of the initial population after mechanical dissociation, but this percentage is 4.4% for enzymatic dissociation (Figure 3-7B, F). These data implies that, relative to the more  55  gentle method of enzymatic dissociation, mechanical dissociation is more detrimental to the viability of hepatoblasts. Alternatively, the observed difference could be attributed directly to the method of dissociation. In enzymatic dissociation, the cell suspension is never filtered through a cell strainer, thereby keeping tissue loss to a minimum. However, in the first step of mechanical dissociation, fetal livers are homogenized and filtered through a 40!m cell strainer. As hepatoblasts often remain as clumps after the initial homogenizing step, a large number of hepatoblasts may fail to pass through the 40!m cell strainer, resulting in mechanically dissociated fetal liver cells to actually be depleted of hepatoblasts. Coupled with the fact that enzymatic dissociation is the more popular method for extracting hepatic cells (Tanimizu et al., 2003; Kamiya et al., 1999), fetal livers were enzymatically dissociated whenever possible in this study. However, due to time and budget constraints, the first two Tag-seq libraries in this thesis study were constructed using fetal livers that were mechanically dissociated (Table 4-1).  56  57  Figure 3-1 DLK1+ cells are present throughout fetal liver development. FACS plots with PE fluorescence on the X-axis and FITC fluorescence on the Y-axis. Top panel shows FACS plots of fixed, unstained fetal liver cells that serve as negative controls for determining FACS gates. Bottom panel shows FACS plots of fixed fetal liver cells and stained with an anti-DLK1 PE-conjugated antibody. Percentage of DLK1+ cells in E10.5, E12.5, E14.5, and E16.5 fetal livers are as indicated, and it is highest in E10.5 fetal livers but progressively drops over the course of embryonic development. The Y-axis shows autofluorescence in the FITC channel as an arbitrary parameter to separate the cell population. This anti-DLK1 antibody (MBL International) is highly specific against the DLK1 antigen, as it has been shown to not react with DLK1-negative HPPL cells (Tanimizu et al., 2004) and Hmga2-/- mutants (K. Jacob, personal communications).  58  Figure 3-2 Expression of HNF4A and DLK1 in the E9.5 liver bud. Deconvolved immunofluorescent images of an E9.5 embryo cross-section stained with antibodies against DLK1 (green) and HNF4A (red), and counterstained with DAPI to visualize nuclei (blue). At 100x magnification, DLK1 expression is detected brightly in the liver bud and faintly in the heart mesenchyme, whereas HNF4A expression is restricted to the liver bud (yellow boxed region). When the liver bud is viewed at 400x magnification, HNF4A-positive hepatoblasts appear as hepatic cords and are already expressing DLK1. Scale bar = 100!m at 100x and 25!m at 400x. 59  Figure 3-3 Expression of DLK1 and HNF4A largely co-localizes during fetal liver development. Deconvolved immunofluorescent images of fetal liver sections stained with antibodies against DLK1 (green) and HNF4A (red), and counterstained with DAPI to visualize nuclei (blue). (A-C) At E10.5, DLK+ cells appear as hepatic cords and already co-express the hepatoblast marker HNF4A. (D-I) At E12.5 and E14.5, DLK1+ cells become more interspersed but still retained HNF4A expression. (J-L) Co-expression of the proteins continued at E16.5 but some DLK1+ cells were negative for HNF4A (white arrows). Scale bar = 50!m. 60  Figure 3-4 DLK1 and CD45 are expressed in mutually exclusive populations. Fetal livers at the indicated stages were sectioned and stained with antibodies against DLK1 (red) and CD45 (green), and counterstained with DAPI to visualize nuclei (blue). Expression of the DLK1 and CD45 proteins did not overlap at the timepoints examined. Scale bar = 50!m.  61  Figure 3-5 DLK1 does not overlap with PECAM expression in the fetal liver. Deconvolved immunofluorescent images of fetal liver sections stained with antibodies against DLK1 (red) and PECAM (green). (A-I) From E10.5 to E14.5, PECAM expression pattern is tubule-like, which resembles the developing vasculature. This pattern is distinct from the rounder, less elongated DLK1-positive cells. (J-L) At E16.5, PECAM expression is restricted to the portal veins and arteries at prenatal stages as previously observed (Takabe et al., 2012). Some co-localization is observed at rounded cells situated away from the PECAM-positive portal vein (yellow arrowheads), but the diffused pattern of PECAM staining is likely non-specific. Scale bar = 50!m. 62  63  Figure 3-6 DLK1 and desmin expression slightly overlaps during fetal liver development. Deconvolved immunofluorescent images of fetal liver sections stained with antibodies against DLK1 (red) and desmin (green), and the embryonic stage at which the images were taken is indicated on the left of each row. Cells positive for desmin appear sickle-shaped, a pattern that is in accordance with previous studies of mouse fetal liver stellate cells (Nitou et al., 2000; Takabe et al., 2012). DLK1 and desmin expression show some degree of overlap at all stages examined (yellow arrows). Images in the inset column show an enlargement of the yellow-boxed area to the left. Scale bar = 50!m.  64  Cell type marker HNF4A PECAM  E9.5 75.2% (113) NP  CD45  NP  desmin  NP  HMGA2  69.0% (174)  E10.5 89.8% (362) 0% (313) 0% (456) 14.0% (171) 37.5% (208)  D12.5 86.2% (353) 0% (267) 0% (304) 10.0% (240) 12.2% (205)  E14.5 90.2% (410) 0% (278) 0% (295) 2.3% (220) 14.4% (243)  E16.5 74.0% (288) 0% * (333) 0% (382) 5.7% (209) NP  Table 3-1 Percentage of DLK1+ fetal liver cells co-expressing various cell type marker at the indicated timepoints. Percentages were generated by manual counting of immunofluorescent images on fetal liver sections co-stained with antibodies against DLK1 and the corresponding protein marker; see Materials and Methods for details on cell counting procedure. Bracketed numbers represent total number of DLK1+ cells analyzed for the respective categories, and discussion of the HMGA2 staining data is reserved for Chapter 5 but is included here for comparison. No statistics were given here as the counts are cumulative sums of counting several images, and the counting was only performed once. * = co-localized staining that is likely non-specific; NP = experiments not performed.  65  Figure 3-7 Dissociation methods for isolating DLK1+ fetal liver cells. E12.5 fetal livers were either mechanically or enzymatically dissociated, stained with an antiDLK1 FITC-conjugated antibody, and then visualized in FACS plots with FITC signal on the X-axis (denoting DLK1-FITC fluorescence) and PE on the Y-axis (to segregate cells for efficient sorting). (A-D) Less than 1% of mechanically dissociated cells stained for DLK1, and when these cells were sorted by flow cytometry and re-analyzed, purity was just below 40%. Upon re-sorting, DLK1+ cells of beyond 90% purity were obtained. (E-G) Under enzymatic dissociation, 4.4% of the initial population stained for DLK1. After just one round of sorting by flow cytometry, at least 90% of the resultant population was already DLK1positive.  66  Chapter 4: Construction and Analysis of liver Tag-seq libraries  4.1  Generation of Tag-seq libraries using sorted DLK1+ cells and perfused adult liver  tissue To examine the gene expression dynamics in developing hepatoblasts, FACS-isolated DLK1+ cells of at least 90% purity were collected from E10.5, E12.5, E14.5, and E16.5 fetal livers. Cells from multiple sorts were pooled to obtain approximately 1!g RNA for Tag-seq library construction (Table 4-1). To capture gene expression in adult hepatocytes, a fifth Tagseq library was constructed using approximately 1g of a PBS-perfused adult liver (Table 4-1), which contains approximately 80% hepatocytes (Si-Tayeb et al., 2010). To ensure that only high quality tag sequences are used for downstream analysis, the Tag-seq data were filtered using the SSOOHE algorithm (sans singletons and one-offs of highly expressed tags; Morrissy, 2010). Specifically, this filtering algorithm eliminates tags that are observed only once, or singletons, which are derived from lowly expressed genes with less than 1 transcript per cell (Romanuik et al., 2009). Tags detected at less than 100 counts that contain a 1-bp mismatch from a second, more highly expressed tag are also removed, as these tags arise due to PCR and other technical artifacts (Morrissy, 2010). As summarized in Table 4-1, each of the five Tag-seq libraries contain at least 3.3 million SSOOHE tags, a sequencing depth where transcript discovery is fully saturated and thus provides true genome-wide coverage of the transcriptome (Morrissy et al., 2009; Wu et al., 2010). To assign SSOOHE tags to curated transcripts, they were mapped to the Mus musculus mm8 assembly of the RefSeq transcript database (Pruitt et al., 2007). The number of tags that map uniquely to a transcript entry ranged from 1.5 to 4.2 million (or 45.1% to  67  55.2% of the tags) in a library, with the remaining tags either mapping to multiple RefSeq transcript entries or not mapping at all. The percentages of uniquely mapping tags are comparable with previous sequencing-based transcriptome libraries (Hou et al., 2007; Morrissy et al., 2009; Siddiqui et al., 2005; Vrljicak et al., 2010) and thus confirm the sequence quality of our Tag-seq libraries. SAGE-based transcriptome analysis techniques can generate tags that map to the same gene due to alternative 3’ ends and internal polyadenylation sites of mRNA transcripts (Malig et al., 2006; Pauws et al., 2001). Collapsing tags that map to the same gene resulted in a number ranging from 5387 to 8376 genes per Tag-seq library. At the time of analysis, the RefSeq mm8 assembly contained 22,186 genes, implying that the expression of 24.2% to 37.8% of mouse genes would be analyzed in this study.  4.1.1  Stage-specific liver genes are detected in the Tag-seq libraries with expected  patterns To assess the quality of the Tag-seq libraries in a biological context, we searched for transcripts derived from genes that have known expression patterns during liver development. First, we searched for genes that are expressed only in fetal hepatoblasts. Dlk1, the gene that codes for the cell surface marker used for isolating fetal hepatoblasts in this study, has been reported to be present in rare cells in the rat adult liver, such as activated oval cells and hepatic stellate cells (Tanimizu et al., 2004; Zhu et al., 2012). However, Dlk1 expression in the normal mouse adult liver has yet to be assessed. In our fetal liver DLK1+ Tag-seq libraries, transcripts from Dlk1 were detected at a peak of 7,545 tags per million (TPM) at E12.5 and a minimum of 4,169 TPM at E16.5 (Figure 4-1A, Table 4-2). Dlk1  68  transcripts were completely absent in the adult, suggesting Dlk1 to be either expressed very weakly or is completely inactive in bulk adult mouse liver cells. Next, we searched for Gpc3, a cell surface protein that is expressed on rat fetal hepatoblasts and is activated in oval cells during liver regeneration (Grozdanov et al., 2006). Gpc3 transcripts were found at high levels in the fetal liver DLK1+ libraries, with a maximum of 2,103 TPM at E10.5 and a minimum of 337 TPM at E14.5 (Figure 4-1A, Table 4-2). Similar to Dlk1, Gpc3 expression was not detected in the adult liver library, suggesting the function of Gpc3 is conserved between the rat and mice liver. Lastly, expression of the fetal liver-specific serum protein Afp was assayed in the Tag-seq libraries. Afp is a serum protein that is specific to embryonic liver cells but becomes activated during hepatocarcinoma (Abelev, 1968). This expression pattern is recapitulated in the Tag-seq libraries, as Afp expression was detected at between 8,853 to 29,857 TPM in the fetal liver DLK1+ libraries but completely absent in the adult (Figure 4-1A, Table 4-2). Overall, these results show that embryonic liver-specific genes can be reliably detected in the Tag-seq libraries. Moreover, the range of expression levels is highly dynamic. For instance, in the E14.5 DLK1+ library, the expression of Dlk1, Gpc3, and Afp is 5083, 337, and 29,857 TPM, respectively. This is the first study to directly compare the transcript levels of hepatoblast markers, and the results indicate that the amount of transcripts produced for a blood serum protein is significantly higher than that for cell surface molecules. Next, expression of genes common to both fetal hepatoblasts and adult hepatocytes were assessed in the Tag-seq libraries. Ttr encodes a thyroid hormone transport protein that is activated beginning at E10 hepatoblasts, and the expression persists in adult hepatocytes (Murakami et al., 1987). In the Tag-seq libraries, Ttr transcripts were stably detected at a  69  range of 20 to 130 TPM (Figure 4-1B and Table 4-2), which supports previous findings by Murakami et al. Similarly, the serum protein Alb gradually increases during fetal liver development (Sellem et al., 1984), and the fetal liver DLK1+ Tag-seq libraries accurately reflect this pattern: Alb transcripts were present at 5,426 TPM at E10.5, increased to 108,448 TPM by E16.5, and stabilized to 47,907 TPM in the adult liver library (Figure 4-1B and Table 4-2). Lastly, the expression pattern of Ubc was analyzed. Ubc encodes an ubiquitin protein that is required for fetal liver growth and general response to cellular stress in hepatocytes (Ryu et al., 2007), but its role and expression level in the adult liver is not well documented. As shown in Figure 4-1B and Table 4-2, Ubc was stably expressed in all five libraries at a range between 475 to 862 TPM. This finding indicates that Ubc may also regulate stress responses in adult hepatocytes. Taken together, our liver Tag-seq libraries are able to detect genes that are expressed in both fetal hepatoblasts and adult hepatocytes. Moreover, temporal fluctuations in the expression level are observed, and these fluctuations may indicate the changing functionality of hepatocytes as they transition from fetal to adult life. Lastly, genes encoding metabolic enzymes that are specifically expressed in mature hepatocytes were analyzed. Tat and Tdo2, two genes involved in tyrosine and tryptophan metabolism and are thus expected to be present only in adult hepatocytes, were detected in the adult liver library at 287 and 678 TPM, respectively. As expected, neither gene was expressed in the fetal liver libraries (Figure 4-1C and Table 4-2). Similarly, G6pc and Pck1, two genes involved in glucose metabolism and expected to be present solely in adult hepatocytes, were respectively expressed at close to 130 and 170 TPM in the adult liver but were either completely absent or present at very low levels in the fetal liver libraries (Figure  70  4-1C and Table 4-2). Specifically, Pck1 was detected at 8 TPM in the E16.5 DLK1+ library, suggesting that DLK1 marks partially differentiated hepatocytes in the late gestation fetal liver. Lastly, transcripts of Arg1, a urea cycle enzyme expressed in functional hepatocytes, were detected at over 600 TPM in the adult liver library (Figure 4-1C and Table 4-2). Interestingly, we found that Arg1 transcripts were not entirely specific to adult liver cells, as they were observed in the E16.5 DLK1+ library at about 250 TPM and at less than 20 TPM in the other three DLK1+ libraries. This is in accordance with a previous study that reported Arg1 expression in perinatal livers (Haraguchi et al., 1987), and may indicate DLK1 to be expressed on some differentiating hepatoblasts at E16.5. Overall, genes associated with terminally differentiated hepatocytes were preferentially detected in the adult liver Tag-seq library. Interestingly, the genes analyzed were similarly expressed at hundreds of TPM in the adult library (Table 4-2), which may indicate the typical expression range for metabolic enzymes.  4.1.2  Genes expressed in other cell types of the fetal liver are detected at low levels or  absent in Tag-seq libraries Analysis from the previous section demonstrated that stage-specific liver genes are observed in our Tag-seq libraries with the expected expression patterns. To further confirm the quality of our libraries, we next sought to determine if the libraries are depleted of genes expressed in non-hepatic fetal liver cell types. Specifically, the Tag-seq libraries were searched for expression of the myeloid progenitors marker CD45 (McCarthy et al., 1985; Uchida et al., 2001), the mesenchyme markers Lhx2, SMA, and desmin (Kolterud et al.,  71  2004; Nitou et al., 2000; Takabe et al., 2012), and the cholangiocyte markers CK19 and Ggt1 (Shiojiri et al., 1991). As shown in Table 4-2 and Figure 4-1D, transcripts for these genes were either absent or detected at very low levels in our fetal liver DLK1+ Tag-seq libraries. Specifically, CD45 expression was completely absent in the E10.5, E12.5, and E14.5 libraries, and was only detected in the E16.5 library at less than 5 TPM. These data provide strong support for the earlier finding in this thesis that showed DLK1 and CD45 cells to be mutually exclusive in the fetal liver (Figure 3-4). For the mesenchyme genes Lhx2, SMA, and desmin, their transcripts were only detected in the E10.5 and E16.5 DLK1+ libraries at levels of 8 TPM or lower (Figure 4-1D). This finding appears to contradict antibody co-staining results from fetal liver sections that showed some overlap of DLK1 and desmin expression throughout development (Figure 3-6 and Table 3-1). More experiments such as qRT-PCR and generating antibody staining images at higher resolution are required to resolve this difference. However, as the expression of other mesenchyme markers (Lhx2 and SMA) were also low in the fetal liver DLK1+ libraries, it does suggest mesenchyme contamination in DLK1+ cells to be minimal. Lastly, for the expression of cholangiocyte markers, CK19 was observed at less than 2 TPM in the E10.5 and E16.5 libraries. However, Ggt1 was expressed at around 4 TPM in the E10.5 and E12.5 DLK1+ libraries, and the expression increased to around 20 TPM at the E14.5 and E16.5 timepoints (Figure 4-1D). Dlk1 mRNA has been previously reported to be absent on cholangiocytes in E17.5 livers (Tanimizu et al., 2004). Thus, the low expression of cholangiocyte markers in fetal liver DLK1+ libraries might be due to small amounts of contaminating cholangiocytes in the sorted DLK1+ samples. Alternatively, residual amounts of DLK1 could still be present on some early cholangiocytes at E16.5.  72  Taken together, the fetal liver DLK1+ libraries are enriched for hepatic transcripts, while transcripts from non-hepatic genes are highly depleted. Caution must be used when analyzing genes with low expression in the Tag-seq libraries, as they could be attributed to genes expressed in contaminating, non-hepatic cells. Nonetheless, the Tag-seq libraries presented in this study provide a reliable resource for studying the transcriptome of fetal hepatoblasts and adult hepatocytes.  4.1.3  Liver transcription factors can be detected in Tag-seq libraries One of the underlying goals of this thesis project was to generate a resource for  capturing the temporal expression patterns of transcription factors (TFs) during fetal liver development. TFs are key proteins that regulate specific phases of hepatoblast differentiation (Battle et al., 2006; Bort et al., 2006; Eferl et al., 1999; Lee et al., 2005), and profiling their fluctuations in developing hepatoblasts could provide clues to the process of hepatoblast maturation. However, due to the generally low expression levels of TFs and contamination from fetal liver blood cells, expression of liver TFs were not reliably detected in previous fetal liver microarray and SAGE studies (Jochheim et al., 2003; Jochheim-Richter et al., 2006; Li et al., 2008; Mouse Atlas Gene Expression project). In fact, qRT-PCR results using whole fetal livers showed Hnf4! and Foxa2 levels to decrease by two orders of magnitude between E11.5 and E13.5 livers (Jochheim et al., 2004), which coincides with the increasing hematopoietic load of the fetal liver (Kurata et al., 1998; Rich and Kubanek, 1979). Thus, the Tag-seq libraries generated using DLK1+ cells isolated during the course of liver development provide a new opportunity to track expression changes of TFs and potentially gain insights into the underlying mechanism of hepatoblast differentiation. For this analysis 73  we focused on TFs that are known to: 1) maintain the hepatic phenotype in fetal hepatoblasts and adult hepatocytes; 2) promote fetal hepatoblast proliferation and survival; and 3) promote cholangiocyte differentiation. First, expression of Foxa2 and Hnf4a were queried in the Tag-seq libraries. Hnf4a is a TF with central roles during liver development. In the fetal liver, it regulates fetal hepatoblast differentiation by modulating the expression of cell adhesion molecules and other hepatocyte differentiation TFs (Li et al., 2000; Parviz et al., 2003; Battle et al. 2006); in the adult liver, gene function is required for bile acid and lipid homeostasis as well as maintaining the epithelial state of fully differentiated hepatocytes (Gonzalez, 2008; Hayhurst et al., 2001; Inoue et al., 2002; Inoue et al., 2004). As shown in Table 4-2 and Figure 4-2, Hnf4a transcripts were found at between 60 – 110 TPM in the Tag-seq libraries. In contrast to a previous report (Jochheim et al., 2004), Hnf4! expression did not drop during the peak of hematopoietic expansion and stayed steady throughout fetal liver development. Interestingly, these tag count numbers are in the same range as Ttr, a secreted protein that is expressed in both fetal and adult liver cells (Table 4-2), suggesting the Tag-seq technique to be ideal for detecting TFs in hepatoblasts. Another TF-encoding gene with key roles during liver development is Foxa2. This gene is required for the initial specification stage of hepatogenesis (Crowe et al., 1999; Lee et al, 2005) and, similar to Hnf4!, is required in adult hepatocytes for maintaining bile acid, glucose, and lipid homeostasis (Bochkis et al., 2008; Wolfrum et al., 2004). As shown in Figure 4-2 and Table 4-2, Foxa2 is detected in all five Tag-seq libraries. The highest expression level was observed in the adult library at just above 100 TPM, and the lowest expression was found in the E16.5 DLK1+ library gene at 13 TPM. As hepatocytes undergo drastic cellular remodeling and proliferation at E16.5 (Greengard et  74  al., 1972), this finding may indicate Foxa2 to be temporarily repressed in prenatal hepatocytes. Next, the Tag-seq libraries were searched for TFs that regulate hepatoblast survival during fetal liver growth: Xbp1, which promotes fetal liver growth by protecting hepatoblasts from apoptosis (Reimold et al., 2000); Ctnnb1, a gene that promotes hepatoblast survival as well as the formation of tight junctions in maturing hepatocytes (Tan et al., 2008); and Foxm1, which promotes mitosis and cholangiocyte lineage in hepatoblasts (Krupczak-Hollis et al., 2004). While all three of these genes have known roles during hepatoblast maturation, their temporal expression patterns have yet to be examined. In our Tag-seq libraries, Xbp1 was detected stably in all four fetal liver timepoints at 30 – 60 TPM, and Xbp1 transcripts were detected at 22 TPM in the adult library (Table 4-2 and Figure 4-2). This finding is consistent with a hepatocyte-specific role for Xbp1 in regulating glucose homeostasis (Lee et al., 2008) and lipogenesis (Zhou et al., 2011). Similar to the temporal expression pattern of Xbp1, Ctnnb1 transcripts were present in all four fetal liver libraries with at least 40 TPM, and the expression persisted through to adult hepatocytes (Table 4-2 and Figure 4-2). This finding suggests a yet uncharacterized role for Ctnnb1 in the adult liver, which warrants refining the analysis of the hepatocyte-specific Ctnnb1 knockout mutants (Sekine et al., 2011). Amongst the three transcription factors, Foxm1 showed the lowest level of expression: its transcripts were observed at E10.5 at only 10 TPM, dropped to about 5 TPM at E14.5 and E16.5, and became barely detectable in the adult library at less than 0.3 TPM. In hepatoblastspecific Foxm1-/- knockout embryos, hepatoblasts are reduced in number and they also fail to differentiate into functional cholangiocytes (Krupczak-Hollis et al., 2004). Pre-differentiation stage hepatoblasts are known to express the cholangiocyte marker CK19 (Germain et al.,  75  1988). Considered together with the finding that CK19 is not expressed in late gestation DLK1+ cells (Tanimizu et al., 2003), the gradual drop of Foxm1 expression in the DLK1+ Tag-seq libraries could indicate that hepatoblasts maintain their bipotential state by losing expression of cholangiocyte genes during development. Overall, TFs that regulate hepatoblast survival were successfully detected in the fetal liver DLK1+ libraries, and their temporal expression patterns may indicate new roles in the adult liver (Ctnnb1) or known roles during hepatoblast differentiation (Foxm1) and in adult hepatocytes (Xbp1). Lastly, temporal expression of TFs driving cholangiocyte differentiation was analyzed in the Tag-seq libraries. Hepatoblasts differentiate starting at E13.5 to generate cholangiocytes and hepatocytes (Germain et al., 1998), and DLK1 has been previously shown to be absent on cholangiocytes at E17.5 (Tanimizu et al., 2003). These findings suggest that cholangiocyte TFs could be expressed higher in the E10.5 and E12.5 DLK1+ libraries (as DLK1 is still marking hepatoblasts) relative to the libraries at E14.5 and E16.5 (when DLK1 marks hepatoblasts and hepatocytes, but not cholangiocytes). For this analysis, genes such as Tbx3, Sall4, Sox9, and Hnf1" that drive cholangiocyte differentiation (Antoniou et al., 2009; Coffinier et al., 2002; Lüdtke et al., 2009; Oikawa et al., 2009; SiTayeb et al, 2010; Suzuki et al., 2008) were investigated. Of the four TFs, three of them (Tbx3, Sall4, and Sox9) showed the highest level of expression in the E10.5 DLK1+ library, with the expression levels ranging from approximately 20 to 60 TPM (Table 4-2 and Figure 4-2). Interestingly, by E16.5, their expression was reduced to at least one-fifth of that observed at E10.5, suggesting that these three TFs are repressed as hepatoblasts mature over time. As for Hnf1", its transcript levels fluctuated at low levels of between 2 – 10 TPM in the DLK1+ Tag-seq libraries. As this expression level in DLK1+ hepatoblasts is lower relative to  76  that observed for the other cholangiocyte TFs, this suggests Hnf1" is a marker for more differentiated cholangiocytes. Taken together, these data implicate that cholangiocyte TFs are expressed at higher levels in pre-differentiation DLK1+ cells. In summary, in our fetal liver DLK1+ libraries, expression of well-characterized liver enriched TFs such as Foxa2 and Hnf4a were robustly detected in maturing hepatoblasts. This is a significant improvement over previous attempts that profiled TF expression levels using cells from whole fetal livers (Jochheim et al., 2003; Jochheim et al., 2004; Jochheim-Richter et al., 2006; Li et al., 2008). As results here suggest a correlation between the function of a liver TF and its temporal expression during hepatoblast differentiation, future studies using the Tag-seq libraries to study the temporal expression profiles for other TF can bring new insights into their roles during in vivo hepatoblast maturation.  4.1.4  Temporal gene expression patterns observed in Tag-seq libraries can be  validated by qRT-PCR The strength of this thesis study lies in the generation of hepatoblast transcriptome libraries over the course of fetal liver development, and these libraries will allow for a temporal analysis of the hepatoblast transcriptome during in vivo maturation. To confirm that the gene expression fluctuations observed in the Tag-seq libraries are reproducible, 15 genes were selected for qRT-PCR experiments to validate their expression patterns in fetal liver DLK1+ cells. The temporal patterns observed between qRT-PCR and Tag-seq were found to be similar for 10 genes, in that the relative expression levels were matching in at least three of the four timepoints (Figure 4-3A, B). In particular, the temporal expression patterns generated by the two techniques were fully validated for four genes: Bambi, Egr1, Hmga2,  77  and Igf2r (Figure 4-3A). Egr1 has a known role in hepatocytes for activating galactokinase, an enzyme necessary for gluconeogenesis in neonates (Yang et al., 2004). While the other three genes have no previously characterized roles during hepatoblast differentiation, their biological functions make them ideal candidates for being involved in the differentiation process. Bambi is a potent inhibitor of the TGF! pathway (Onichtchouk et al., 1999), a signaling cascade that has been shown to regulate bile duct formation (Ader et al., 2006; Clotman et al., 2005). Hmga2 is highly expressed in embryonic stem cells (Li et al., 2006), and the gene has been shown to regulate renewal of mouse adult neuronal stem cells (Nishino et al., 2008). Igf2r, a “dominant negative” receptor for IGF2 (Hawkes and Kar, 2004), is expressed in some fetal liver cells and can affect the expansion of hematopoietic stem cells by sequestering the IGF2 ligand (Zhang et al., 2004). Overall, these results show that collectively, the DLK1+ Tag-seq libraries is a useful tool for identifying differentially expressed genes in developing hepatoblasts.  4.2  Clustering analysis of Tag-seq libraries reveals distinctive gene expression profiles  in developing hepatoblasts After analyzing our Tag-seq libraries on a gene-by-gene basis, the libraries were subjected to analysis at the genome-wide level. During organogenesis, the underlying gene expression changes are controlled by TFs, and these TFs often regulate a large number of genes to elicit their effects. For instance, HNF4A regulates the expression of more than 60 genes during hepatoblast differentiation (Battle et al., 2006; Li and Duncan, 2001). This suggests that genes with a similar function are often co-regulated and thus co-expressed during development. This concept of has been applied in a wide spectrum of studies to 78  generate novel gene-function associations based on shared temporal expression patterns (Sásik et al., 2002; Stevenson et al., 2003; Wamstad et al., 2012). Thus, by grouping together genes with a similar temporal expression pattern in our liver Tag-seq libraries, novel factors that drive specific phases of liver development may be identified. To segregate genes based on their temporal expression pattern, Tag-seq libraries were analyzed using K-means clustering (Cai et al., 2004). The algorithm requires the user to supply a dataset and a K value, which designates the total number of clusters to be generated. Based on these inputs, the algorithm divides the data points into K clusters, with each cluster consisting of data points that share a stereotypical pattern within the particular dataset. Kmeans clustering is ideal for count-based data such as SAGE libraries (Cai et al., 2004), and the algorithm has been widely used to extrapolate gene expression signatures from sequencing-based transcriptome data (Blackshaw et al., 2004; Hoffman et al., 2008; Vrlijcak et al., 2010). To prepare our Tag-seq data for clustering, counts from tags belonging to the same gene were summed to generate a number representing the gene’s total expression in a library (see Materials and methods), and lowly expressed genes were removed to reduce noise in downstream analysis. The majority of hepatic genes were detected in our Tag-seq libraries with at least 50 TPM (such as the TFs Xbp1 and Foxa2; Table 4-2). The liver gene with the lowest expression, Foxm1, was found at approximately 10 TPM. In contrast, non-hepatic transcripts were present at much lower levels, such as the mesenchyme marker Lhx2, myeloid progenitors marker CD45, and cholangiocyte marker CK19 (fewer than 8, 5, and 2 TPM in the fetal libraries, respectively; Table 4-2 and Figure 4-1). Thus, we placed a  79  threshold at 10 TPM where, in order for a gene to be included for clustering analysis, it had to be expressed at more than 10 TPM in at least one of the five Tag-seq libraries. Applying this threshold, a list of 5449 genes was obtained. Subjecting these genes to the K-means algorithm (Cai et al., 2004) yielded 14 distinct clusters (Figure 4-4). Visual inspection of the 14 gene clusters revealed four general patterns, with the peak of expression occurring either at the E10.5 (clusters A, B, C), E12.5 or E14.5 (clusters D, F, G, H), E16.5 (J, K, L), or adult timepoint (M, N). Note that the temporal expression patterns for two clusters (E and I) do not show a single, distinct peak. In this study, cluster I will be classified as bimodal (with expression peaks at both the E10.5 and E16.5 timepoints), and cluster E will be classified as constant (with slight peaks at the E12.5 and adult timepoints). To determine if K-means clustering was successful in generating biologically significant gene clusters, genes within each cluster were inspected in detail. Genes with similar functions during development could be co-regulated and thus share the same temporal expression pattern (Allocco et al., 2005). To investigate this in the context of liver development, the clusters generated from K-means were searched for genes that 1) encode key transcription factors and serum proteins expressed in hepatic cells; or 2) are central components of signaling pathways that have been implicated to regulate liver development. In total, 80 genes were identified, of which 35 had no direct literature support in regulating liver development. The 45 remaining genes were divided into markers of hepatoblasts, regulators of hepatoblast proliferation/differentiation, or factors associated with the fully differentiated phenotype of hepatocytes, as shown in Table 4-3. Here, clusters with a similar temporal expression pattern do contain genes that regulate the same aspect of liver development. For genes regulating hepatoblast proliferation, 17 genes were identified, and  80  they were distributed in clusters that peak at E10.5 (Smad2, Tgfb2, Dlc1, Hex, Onecut2, Igf2bp1), E12.5/14.5 (Ctnnb1, Igf2r, Insr, Max), or E16.5 (H19, Id3, Igfbp1, Igfbp3, Jun). Interestingly, hepatoblast proliferation genes were absent in the bimodal (cluster I) and constant cluster (cluster E). This may indicate that genes controlling cell cycling of hepatoblasts are stably expressed throughout fetal liver development, and these genes are silenced in quiescent hepatocytes of the adult liver. For hepatoblast markers, 4 out of 6 (Dlk1, Afp, Cdh1/E-cadherin, Met) were grouped together in clusters with peak expression at E12.5 or E14.5, suggesting expression of hepatoblasts markers to peak after the initial phase of liver bud expansion but drops soon after hepatoblasts begin to differentiate. In contrast, clusters with E12.5/14.5 expression peaks contain only 2 out of 14 genes (Ctnnb1, Tgfbr2) with a role in hepatoblast differentiation. This finding implies that clusters with a temporal expression peak at E12.5/14.5 may enrich for genes associated with hepatoblast markers but not hepatoblast differentiation. Lastly, genes belonging to clusters with constant or specifically high expression at the E16.5 or adult timepoint contain 17 of the 23 genes associated with fully differentiated hepatocytes, which correlates well with the appearance of hepatocytes at late gestation (Germain et al., 1988; Shiojiri, 1981; 1994). Overall, K-means clustering analysis was able to group together genes with similar liver functions based on their temporal expression patterns in the Tag-seq libraries.  4.2.1  Temporally distinct gene clusters are enriched for different Gene Ontology  terms Results from the previous section suggest that K-means clustering can group together genes with similar biological functions. To extend those findings globally, all clusters were  81  analyzed for the enrichment of Gene Ontology (GO) terms, which are used to functionally categorize genes based on biological process, molecular function and cellular component (Ashburner et al., 2000). GO enrichment was analyzed using DAVID, an online tool that discovers shared Gene Ontology categories in a user-specified gene set (Huang et al., 2007). Specifically, each cluster was individually loaded into DAVID and analyzed for GO enrichment relative to the 5449 genes used for K-means analysis. As expected, clusters with peak expression in the adult liver (clusters M and N) are enriched for GO terms ‘carboxylic metabolic process’ and ‘lipid metabolic process’ (Table 4-4). This enrichment is reflective of the metabolic functions in adult hepatocytes and is highly specific to clusters M and N (Figure 4-5). The E16.5-peaking cluster L showed a high enrichment for ‘chemotaxis’ (Table 4-4 and Figure 4-5). Interestingly, this finding was highly specific to cluster L, as the GO term was not significantly enriched in other clusters with E16.5 expression peaks (clusters J, K, and the bimodal cluster I; Table 4-4 and Figure 4-5). Similarly, the E14.5-peaking cluster F was significantly enriched for ‘wound healing’, but such enrichment was not observed for clusters with similar expression patterns (clusters D, G, H, and the constant cluster E; Table 4-4 and Figure 4-5). Lastly, clusters with peak expression in the E10.5 DLK1+ library (clusters A, B, C) are all enriched for ‘gene expression’ (Table 4-4), and the enrichment is highly specific to clusters with this temporal expression pattern (Figure 4-5). As many transcription factors are classified under the ‘gene expression’ category, this finding implies E10.5 DLK1+ cells to be specifically enriched for transcription factors relative to hepatoblasts of later stages. Significantly, this is a novel finding that was not discovered by previous fetal liver transcriptome studies (Jochheim-Richter et al., 2006; Li et al., 2008; Otu et al., 2007) and thus may provide new clues to defining early hepatoblasts.  82  4.3 4.3.1  Discussion Tag-seq libraries accurately captured the changing transcriptome during liver  development In this chapter, I presented the construction and analysis of five Tag-seq libraries using fetal hepatoblasts and adult hepatocytes. The Tag-seq libraries were highly and specifically enriched for hepatic transcripts while those from non-hepatic cells were depleted. For this study, SSOOHE tags from the Tag-seq libraries were mapped to the RefSeq database (Pruitt et al., 2007), which is a collection of sequences that is curated based on known mRNA sequences. As this study is concentrated on the transcriptome, RefSeq was chosen as the database of choice for performing tag-to-gene mapping. Overall, 49.2% of SSOOHE tags in the five Tag-seq libraries mapped uniquely to RefSeq transcripts (Table 4-1). This implies that just over 50% of SSOOHE tags were discarded from further analysis. A portion of these tags mapped to the RefSeq database but did so either ambiguously or to an antisense strand. The remaining tags are likely to map to other sequence databases such as Ensembl (Birney et al., 2004) and the Mammalian Gene Collection (Gerhard et al., 2004), which has been observed for LongSAGE and Tag-seq libraries (Hou et al., 2007; Morrissy, 2010; Vrljicak, 2010). Lastly, some tags will map to the mouse genome, which could represent novel transcripts. At the time of analysis, the RefSeq mm8 database was the most current assembly and was thus used for mapping of the tag sequences. Since then, two new assemblies (mm9 and mm10) have been completed. While the newer assemblies contain up-to-date genome annotations such as new genomic coordinates and transcript isoforms, the gene symbols for  83  well-characterized genes remain unchanged relative to older assemblies. As gene symbols were used in this study as unique identifiers for each tag sequence, using RefSeq mm8 for mapping purposes can be justified. Notably, numerous web-based bioinformatics tools still allow users to input mm8 data (Akagi et al., 2010; Giardine et al., 2005; Lan et al, 2011), suggesting mm8 to be a version of the mouse genome assembly that is still useful at large. However, for future studies utilizing Tag-seq data from this study, it is advisable to remap tag sequences to more recent RefSeq assemblies such as mm9 or mm10 in order to capture the most updated genome annotations. One of the purposes for using Tag-seq to analyze the hepatic transcriptome was to follow the temporal expression of transcription factors. To our knowledge, this is the first time that the expression of hepatic transcription factors during in vivo hepatoblast maturation was successfully captured in a temporal transcriptome study. From the Tag-seq libraries, two observations were made for liver TFs in the Tag-seq libaries. For instance, in the E14.5 DLK1+ library, Foxa2 and Hnf4! (TFs that play central roles in the homeostatic ability of adult hepatocytes and maintenance of the hepatocyte phenotype) are readily detectable at 64 and 91 TPM, respectively; Ctnnb1 and Xbp1 (TFs with roles in promoting hepatoblast survival) are detected at a slightly lower level of 33 and 40 TPM, respectively; and Sox9, Hnf1", Foxm1, Tbx3, and Sall4 (TFs driving cholangiocyte differentiation) are detected at even lower levels of 0, 10, 5, 18, and 12 TPM, respectively (Table 4-2). Therefore, a correlation exists between the role of a liver TF and its expression level in the E14.5 DLK1+ library. The second pattern is observed within the E16.5 DLK1+ library and involves the TFs maintaining hepatocyte phenotype and hepatoblast proliferation. Here, Foxa2 and Hnf4! show a drop in expression levels relative to E14.5. In contrast, expression levels of Ctnnb1  84  and Xbp1 increased almost by two-fold relative to the E14.5 timepoint (Table 4-2). This finding may indicate that as hepatocytes enter the proliferation phase at E16.5 (Greengard et al., 1972), genes promoting hepatoblast division are temporarily upregulated while those controlling hepatocyte identity are repressed. Interestingly, a role for HNF4A to suppress proliferation in adult hepatocytes was recently discovered (Bonzo et al., 2012). Further qRTPCR experiments using E14.5 and D16.5 DLK1+ sorted cells can be performed to confirm the observed expression patterns for Foxa2, Hnf4!, Ctnnb1, Xbp1, as well as other genes involved in regulating the hepatocyte phenotype such as Cited1, Hnf1!, Pxr (Li et al., 2000; Qu et al., 2007) and hepatocyte proliferation such as Jun and Id3 (Eferl et al., 1999; Hilberg et al., 1993; Nakayama et al., 2006). For 15 genes, their temporal expression patterns as observed in DLK1+ Tag-seq libraries were independently assessed by qRT-PCR, a technique that is considered as the gold standard for quantifying mRNA quantity due to its sensitivity as well as the lack of amplification and processing of the cDNA (Wong and Medrano, 2005). In this study, qRTPCR was able to fully duplicate the temporal patterns for 4 of the 15 genes (at 4 out of 4 timepoints; Figure 4-3A), with 6 additional genes whose patterns were duplicated partially (3 out of 4 timepoints; Figure 4-3B). Incomplete correlation of the expression values between Tag-seq and qRT-PCR is not unexpected, as the success rate of using qRT-PCR to validate differentially expressed genes in transcriptome studies can be as low as 50% (Anisimov et al., 2002; Garbett et al., 2008; Millien et al., 2008). Data discrepancy between the two techniques could be attributed to many factors. First, DLK1+ cells for Tag-seq library construction and qRT-PCR validation were collected from separate days, and the use of nonidentical starting material may result in slight variations in gene expression. Moreover,  85  differences in experimental procedures such as the round of PCR amplification during construction of transcriptome libraries could also play a role (Duggan et al., 1999; Velculescu et al., 1995). Furthermore, in this study, cDNA used for Tag-seq library construction was generated with oligo-d(T) primers, whereas cDNA used for qRT-PCR was generated with random hexamers, which could result in differences in the final pool of cDNA material. In addition, different methods of data normalization could also lead to differences in determining gene expression levels. qRT-PCR results are normalized against the expression of a single housekeeping gene, in this case !-actin. In contrast, Tag-seq data is normalized against all high-quality tags, a method that is routinely used for sequencing-based libraries (Li et al., 2010; Vrljicak, 2010; Yang et al., 2010). Other methods of normalizing countbased data have been described. One such method normalizes expression against multiple housekeeping genes, which has been shown to yield normalized values that are more comparable to qRT-PCR results (Lu et al., 2009). More recently, a normalization method was designed to specifically offset the bias caused by the domination of ultra highly expressed genes (Robinson and Oshlack, 2010). This method could be relevant to our Tag-seq libraries, as transcripts for genes such as Dlk1, Afp, and Alb are detected at several orders of magnitude above other genes with known roles in liver development (Table 4-2). Lastly, the SYBR Green technology used in this study for qRT-PCR validations may not be sensitive enough to detect gene expression changes as observed in Tag-seq. Thus, using more sensitive real-time assays such as TaqMan probes (Kuimelis et al., 1997) or molecular beacons (Manganelli et al., 2001) may allow small gene expression fluctuations to be detected and thereby improve current validation rates. Overall, validation results presented here highlight the need to  86  independently confirm the differential gene expression of candidate genes (either at the mRNA level through qRT-PCR/in situ hybridization or at the protein level through antibody protein staining/Western blots) before embarking on further functional analyses.  4.3.2  Temporal gene clusters may yield novel, stage-specific insights for hepatoblasts In summary, K-means clustering generated 14 gene clusters with distinct temporal  expression patterns during liver development (Figure 4-4), and genes sharing similar temporal expression patterns were found to have similar functions during liver development (Table 4-3). As expected, genes with highly specific expression in the adult liver Tag-seq library are involved in the metabolic functions of fully differentiated hepatocytes (clusters M and N). Also, Foxa2 and Hnf4!, two genes with key functions in both fetal hepatoblasts and adult hepatocytes are found in cluster E, which shows stable expression during the course of liver development. More importantly, clusters with gene expression that peak at different stages of embryogenesis were also identified. Specifically, clusters with expression peaks at the E10.5 and E16.5 timepoints are enriched for genes known to regulate hepatoblast proliferation and differentiation, which is in accordance with the hepatogenesis events occurring at those stages. Interestingly, some of these clusters contain sharp peaks (ie. cluster A, I, and L), suggesting the corresponding genes to be tightly regulated during hepatoblast maturation. Lastly, most hepatoblast-specific cell surface markers are distributed across clusters with peak expression at E12.5 or E14.5. In contrast to clusters with peak expression at E12.5 or E16.5, clusters with this type of expression pattern contain relatively duller peaks (ie. cluster D and F), suggesting the corresponding genes to be more stably expressed during the course of fetal liver development.  87  Amongst the 14 gene clusters generated in this study, several clusters warrant further analysis. For instance, cluster A has a distinctive temporal expression pattern during hepatoblast maturation: a high peak in E10.5 DLK1+ cells, dropping sharply to a steady level between E12.5 and E16.5, and finally minimal expression is reached in the adult (Figure 4-4). Expression of cholangiocyte and hepatocyte markers in the fetal liver is first observed around E13.5 (Shiojiri, 1981; 1994). Coupled with the sharp decrease in expression for cluster A genes between the E10.5 and E12.5 timepoints, this indicates that cluster A genes become repressed as early hepatoblasts prepare for differentiation. This notion is supported by the fact that cluster A contains genes that specifically regulate the early stages of hepatoblast maturation such as Smad2 (Weinstein et al., 2001) and Tgfb2 (Stenvers et al., 2003), as well as genes expressed in stem or undifferentiated progenitor cells such as Lin28 (Yokoyama et al., 2008; Zheng et al., 2009) and Hmga2 (Li et al., 2006; Nishino et al., 2008). Another cluster of potential interest is cluster L. In addition to the cluster having a very distinctive peak at E16.5, it also contains only 91 genes, the fewest number amongst the gene clusters generated from K-means analysis. Genes belonging to cluster L include Jun and Junb (expressed at 350 and 65 TPM in the E16.5 DLK1+ library, respectively), and the JUN proteins have been shown to regulate hepatogenesis by maintaining hepatoblast proliferation and integrity in the E11.5 to E13.5 liver (Eferl et al., 1999; Hilberg et al., 1993). Interestingly, they are also activated during the initial phase of liver regeneration to induce cell cycle entry in quiescent hepatocytes (Fausto, 2000). However, the role of JUN proteins in prenatal hepatoblasts has yet to be characterized. To explore this further, the expression of Jun in E16.5 DLK1+ cells need to be first confirmed at both the transcript and protein levels.  88  Transgenic animals with hepatoblast-specific Cre expression driven by the Alb-promoter Afp-enhancer element (Kellendonk et al., 2000; Parviz et al., 2003) could be mated to animals carrying the floxed allele of Jun (Palmada et al., 2002) to generate a conditional knockout for studying the effects of the JUN proteins during the development of pre-natal hepatocytes. Furthermore, many genes in cluster L are cytokines, such as Cxcl1, Cxcl2, Cxcl10, and Cxcl16 (expressed at 85, 111, 47, and 12 TPM in the E16.5 DLK1+ library, respectively). At this point, while the roles that these cytokines play in the late gestation fetal liver are still unclear, they may be involved in the trafficking of hematopoietic cells during development. During embryogenesis, hematopoietic progenitors home and expand in the fetal liver environment to generate various blood cell lineages (Crawford et al., 2010), a process that requires the concerted action of various cytokines and their receptors (Ma et al., 1999; Crawford et al., 2010). SCF+ DLK1+ cells in the E15.5 fetal liver have been shown to produce several cytokines that support the maintenance and expansion of hematopoietic stem cells, with one of those cytokines being a CXCL-type protein (Chou and Lodish, 2010). Also, at E17.5, the granulocyte and B lymphocyte populations expand in the fetal liver (Crawford et al., 2010). As hepatoblasts have been shown to regulate B lymphopoiesis in the fetal liver (Bouzin et al., 2003), data from the Tag-seq libraries suggest that late gestation hepatoblasts may regulate granulocyte and lymphocyte development via the secretion of cytokines. Interestingly, chemokines may also have a negative impact on hematopoietic cell number in the fetal liver. In the late gestation fetal liver, hematopoietic cells begin to exit the organ and home into the bone marrow (Crawford et al., 2000). In the adult bone marrow, hematopoietic cells exit the bone marrow niche as part of homeostasis, and this egression is signaled by  89  metalloproteinase-mediated cleavage of CXCL-type chemokines (Lévesque et al., 2003; Van Lint and Libert, 2007; Vagima et al., 2011). Notably, genes encoding metalloproteinases such as Mmp2 and Mmp23 are also found in cluster L (see Appendix L). These data suggest that proteolytic cleavage of chemokines may also underlie hematopoietic egression during in the embryonic liver. To investigate this further, protein expression of the chemokines as well as their corresponding cleaved products would need to be confirmed in E16.5 fetal livers. Also, E16.5 DLK1+ cells can be co-cultured in vitro with hematopoietic stem cells to test their ability to hinder stem cell expansion. In addition, knockout mice for the chemokines expressed highly in E16.5 DLK1+ cells could be assayed for deficiencies in the trafficking of hematopoietic cells in perinatal embryos. Lastly, clusters containing hepatoblast cell surface molecules could be further analyzed. Various markers for fetal hepatoblast have been identified to date: DLK1 (Tanimizu et al., 2003), E-cadherin (Nierhoff et al., 2005), MET (Suzuki et al., 2002), and GPC3 (Grozdanov et al., 2006). Interestingly, the markers were segregated into different temporal clusters: DLK1 is in cluster D (peak expression at E12.5), MET and E-cadherin are in cluster F (peak at E14), and GPC3 in cluster I (bimodal peaks at E10/E16). If hepatoblasts were a homogenous population of cells, then the various hepatoblast cell surface markers would be expected to be grouped under clusters with similar temporal expression patterns. However, as the hepatoblasts markers are segregated into different clusters, this finding may suggest that hepatoblasts are a heterogeneous population. This is supported by a report by Tanaka et al., where the in vitro proliferation capacity of DLK1+ cells can be determined by the presence or absence of the epithelial cell adhesion molecule EpCAM (Tanaka et al., 2009). Thus, our K-means clustering results suggest a need to thoroughly analyze fetal liver 90  subpopulations that express some combination of the DLK1/GPC3/E-cadherin/MET proteins, as the individual subpopulations may represent unique subsets of hepatoblasts. Specifically, the subpopulations can be isolated and assayed for differentially expression of hepatocytes versus cholangiocyte markers, proliferation and differentiation potentials in primary cultures, as well as their ability to repopulate an injured liver upon transplantation. Furthermore, as hepatoblast markers are found in clusters D, F, and I, deeper analysis of these clusters could provide insights into genes and pathways that regulate the identity and differentiation potential of fetal liver progenitors. As a means to perform a global analysis of the gene clusters, Gene Ontology enrichment analysis was performed on the 14 gene clusters. Certain GO terms were found to be significantly and specifically enriched in some clusters (Table 4-4 and Figure 4-5). Care must be taken when interpreting results from GO enrichment analysis. For instance, while some clusters with peak expression at E12.5/E14.5 or E16.5 enrich for distinct GO terms (‘chemotaxis’ in E16.5-peaking cluster L, and ‘wound healing’ in E14.5-peaking cluster F), the same terms were not enriched in other clusters that share their respective temporal expression patterns. While this could indicate a highly specific function for genes within clusters L and F, the lack of support shown by other clusters with similar expression patterns could also indicate the finding to be a false-positive. Furthermore, GO enrichment analysis is often dominated by a few genes that are at the node of several signaling pathways (Gillis and Pavlidis, 2011). By definition, these genes are associated with multiple GO terms and thus possess high multifunctionality, which could dramatically skew the assessment of GO term enrichment within a given gene set. Therefore, curating gene clusters to exclude genes with  91  high multifunctionality before proceeding to GO analysis could generate more dependable GO enrichment results. In addition to identifying shared biological functions, another common method for analyzing gene sets is to identify shared cis-regulatory elements. Transcription factors are central to the regulatory network underlying both fetal and adult liver development (Kyrmizi et al., 2006; Odom et al., 2004). The DNA motifs recognized by different transcription factors have been categorized in databases such as JASPAR and TRANSFAC (Sandelin et al., 2004; Wingender et al., 2000). Therefore, genes in the same cluster can be queried for shared regulatory elements and enriched transcription factor binding sites (TFBS) to gain insights in the transcriptional regulatory networks that are active during the course of in vivo hepatoblast maturation. Various bioinformatics tools are available for discovering shared transcription factor binding sites within the cis-regulatory regions of co-expressed genes. For instance, DiRE focuses the search of enriched TFBS within promoters as well as distant enhancers that are evolutionarily conserved across mammalian genomes (Gotea and Ovcharenko, 2008). oPOSSUM is another analysis algorithm that identifies enriched binding sites based on the number of TFBS and their proximity to each other within the regulatory elements of user-specified genes (Ho Sui et al., 2005; Kwon et al., 2012). While this algorithm may miss TFBS that are enriched in distant enhancers, its added emphasis on the number and proximity of TFBSs may yield in silico predictions with improved biologically significance.  92  Library Litters Method of (embryos) diss’n  Cells collected (x1000)  Total tags in SSOOHE dataset  E10.5 DLK1+ E12.5 DLK1+ E14.5 DLK1+ E16.5 DLK1+ Adult liver  23 (174)  Enzymatic  262  8,353,961  4 (35)  Mechanical  125  4,474,000  3 (24)  Mechanical  137  7,469,787  6 (9)  Enzymatic  232  7,649,649  n/a (1 perfused liver)  Perfused  n/a (~1g tissue)  3,356,883  Tags uniquely mapping to mm8 transcripts (%) 3,816,645 (45.7%) 2,209,322 (49.4%) 3,770,756 (50.5%) 4,221,576 (55.2%) 1,513,829 (45.1%)  Total tag types  45,780 38,408 40,306 48,269 34,302  Tag types uniquely mapping to mm8 transcripts (%) 22,026 (48.1%) 14,378 (37.4%) 16,707 (41.5%) 22,407 (46.4%) 9,231 (26.9%)  Genes after collapsing  8,082 6,605 7,213 8,376 5,387  Table 4-1 Summary of Tag-seq library constructed. Total tags represent the number of Tag-seq tags that were sequenced per library and retained within the SSOOHE (sans singletons, one-offs of high expresser) dataset; numbers in this column is representative of library depth. These tags were mapped using Discovery Space to the mouse RefSeq transcript database (mm8), and tags that mapped unambiguously to a single entry were used to generate values for the Tags uniquely mapping to mouse RefSeq (%) column. Total tag types represent the total number of tags with unique sequences. Tag types uniquely mapping to mouse RefSeq (%) represent tag types that mapped unambiguously to RefSeq mm8. Genes after collapsing represent the number of genes that are represented by all uniquely mapping tag types.  93  94  Figure 4-1 Tag-seq libraries faithfully capture expression of stage-specific liver genes. X-axis indicates the five Tag-seq libraries that were generated, and Y-axis depicts the abundance of tags that map to the particular gene, normalized to tags per million. Thick grey overlay line denotes normalized expression at 10 TPM. The Tag-seq libraries constructed contain tag types from genes with known liver functions. (A) Tag types from embryonic-specific genes were only present in the embryonic libraries, but completely absent in the adult library. Expression of Dlk1 is red, Gpc3 in blue, and Afp in green. (B) Tag types from genes that are required for fetal as well as adult liver cell functions were found in all five Tag-seq libraries. Alb is in green, Ttr in blue, and Ubc in red. (C) Tag types for genes such as metabolic enzymes that are specifically expressed in the adult liver were detected in high abundance only in the adult library. Tat in yellow, Tdo2 in black, G6pc in blue, Pck1 in red, and Arg1 in green. (D) Tag types mapping to genes that are markers of other cell types were present, but at very low levels of less than 10 tags per million. CD45 in blue, Lhx2 in yellow, SMA in orange, desmin in red, Ggt1 in dark green, and CK19 in light green.  95  Fetal only Fetal and Adult  Adult only  Other cell types  Liver TFs  Gene Dlk1 Gpc3 Afp Alb Ttr Ubc Tat Tdo2 G6pc Pck1 Arg1 CD45 Lhx2 Des Sma Sox9 Hnf1b Foxa2 Hnf4! Xbp1 Ctnnb1 Foxm1 Tbx3 Sall4  E10.5 DLK+ 6093 2103 8853 5426 66 596 0 0 0 0 16 0 7 4 3 20 8 37 85 40 144 10 45 58  E12.5 DLK+ 7545 670 14794 48823 20 589 0 0 0 0 10 0 0 0 0 0 5 60 113 60 193 8 35 20  E14.5 DLK+ 5082 337 29857 47360 47 475 0 0 0 0 12 0 0 1 0 0 10 64 91 33 40 5 18 12  E16.5 DLK+ 4169 2026 24624 108448 132 862 0 0 1 8 252 3 3 6 3 4 3 13 61 53 62 5 6 8  Adult liver 0 0 0 47907 74 835 287 678 130 170 611 4 0 0 2 0 18 102 103 22 45 0 24 1  Table 4-2 Transcript abundance of genes of interest as detected in Tag-seq libraries. Numbers depict gene expression level normalized to tags per million (TPM). Overall, serum proteins (Afp and Alb) are the most abundant, followed by cell surface molecules (Dlk1 and Gpc3) and transcription factors (Foxa2 and Hnf4a). If a gene is undetected or is present at less than 5 raw tags in a library, expression is counted as zero.  96  Figure 4-2 Temporal expression patterns of hepatoblast/cholangiocyte TFs in Tag-seq libraries. Expression levels (shown on Y-axis, in tags per million) of TFs with known roles during hepatoblast differentiation as observed in the respective Tag-seq libraries (shown on X-axis). TFs with known roles in both hepatoblasts and hepatocytes are shown as black lines, while those regulating hepatoblast survival is in red and TFs regulating cholangiocyte differentiation are shown in light blue. Thick grey overlay line denotes normalized expression at 10 TPM.  97  98  Figure 4-3 qRT-PCR validation of 15 genes and their temporal expression patterns in DLK1+ Tag-seq libraries. X-axis denotes developmental timepoints at which DLK1+ cells were collected; left Y-axis shows gene expression in Tag-seq libraries (in TPM), and right Y-axis shows expression level by qRT-PCR (means from two biological replicates, with error bars indicating standard deviation). Genes with matching Tag-seq and qRT-PCR expression levels at all four timepoints are shown in panel A, whereas genes with matching expression levels at three, two, or one timepoint are shown in panels B, C, and D, respectively.  99  Figure 4-4 Fourteen temporal clusters were generated from K-means clustering. Each graph shows the temporal expression of genes grouped into the indicated cluster by the K-means clustering algorithm (Cai et al., 2004). X-axis shows the timepoints of liver cells used for Tag-seq library construction; Y-axis shows relative gene expression level, which is determined by taking a gene’s normalized expression at a given timepoint (in TPM) and dividing by the cumulative expression of that gene across the five Tag-seq libraries. Thus, a relative expression of 1 indicates the gene to be only expressed in the corresponding Tag-seq library, whereas a value of 0 indicates the gene to be completely silence. Each line in a graph depicts the temporal expression pattern of one gene. In total, fourteen gene clusters were discovered, with each cluster having a distinct temporal pattern. See Appendix for complete list of genes in each cluster.  100  Peak expression  Cluster (# genes)  A (179)  E10.5  B (781)  C (724)  E12.5/14.5  D (326)  Constant  E (570)  F (331)  E12.5/14.5 G (171)  H (518)  Bimodal  I (331)  Genes Dnmt1 Fzd2 Hmga2 Lin28 Smad2 Tgfb2 Yap1 Dlc1 Dvl2 Hex Lats1 Onecut2 Pik3c2a Pten Tgfbr1 Cited1 Dnmt3a Hnf1a Igf2bp1 Igf2bp2 Igf2bp3 Myc Mycn Cdkn3 Ctnnb1 Dlk1 Dicer1 Igf2r Cebpg Foxa2 Foxo1 Hnf4a Ide Afp Cdh1 Insr Met Mst1 Tgfbr2 Gsn Ilk Itga2b Lin28b Nomo1 Aph1a Mapk1 Max Pknox1 Cebpa Gpc3 Hes1  Hepatoblast marker  Hepatoblast Proliferation  Hepatoblast Differentiation  Differentiated hepatocyte  101  Bimodal  I (331)  J (507)  E16.5  K (288)  L (95)  M (431) adult N (197)  Id1 Id2 Tgif1 Ets2 H19 Hes6 Jund Meis1 Alb Apoe Cebpd G0s2 Id3 Igfbp1 Igfbp3 Insig1 Ttr Cebpb Egr1 Egr2 Jun Junb Apoa2 Dvl1 Egfr Gck Insig2 Rxrg Tat  Table 4-3 Regulatory genes are distributed among the 14 gene clusters. Each cluster was systematically searched for genes with known roles during liver development or central roles in key signaling pathways. A list of 80 genes was generated and searched in the literature for direct evidence as a hepatoblast marker (green box), regulator of hepatoblast proliferation (blue box) and differentiation (red box), and marker of the terminally differentiated hepatocytes (brown box). If literature support is present, the corresponding box was shaded with the assigned colour.  102  Gene Ontology term 0010467: gene expression 0006397: mRNA processing  0042060: wound healing  0019538: protein metabolic process  0006935: chemotaxis 0019752: carboxylic acid metabolic process 0006629: lipid metabolic process  Cluster A B C A B C D E F G H D E F G H I J K L M N M N  Expression peak in Tag-seq libraries  E10.5  E12.5/E14.5  E16.5  Adult  Count (% cluster) 58 (34.5) 171 (23.9) 194 (28.5) 7 (4.2) 20 (2.8) 32 (4.7) 0 0 12 (3.9) 4 (2.5) 0 70 (22.6) 119 (22.2) 62 (20.0) 34 (21.1) 131 (16.7) 0 4 (0.9) 19 (3.7) 10 (11.1) 47 (11.6) 18 (9.7) 45 (11.1) 28 (15.1)  Fold enrichment 2.16 1.52 1.89 3.52 2.74 4.21 n/a n/a 8.14 5.29 n/a 1.49 1.44 1.184 1.27 1.70 n/a 1.49 6.25 17.91 4.55 3.65 3.36 4.38  Bonferroni 4.90 E-06 7.49 E-06 4.00 E-18 1 1.08 E-01 1.30 E-07 n/a n/a 1.21 E-03 1 n/a 7.21 E-03 6.28 E-02 1 1 2.42 E-07 n/a 1 1.49 E-01 1.93 E-05 3.83 E-14 4.09 E-02 1.36 E-08 7.67 E-07  Table 4-4 Top enriched GO terms based on clusters with similar temporal patterns. Select Gene Ontology terms with a fold enrichment of at least 1.45 over the background gene set are listed. Clusters with a similar temporal expression pattern are shown together, with exceptions for cluster E (constant expression) and cluster I (bimodal peaks) which were arbitrarily placed in the table for comparison. Bonferroni values are p-values corrected for multiple hypothesis testing, and values smaller than 0.05 was considered significant enrichment.  103  Figure 4-5 Genes with preferential expression in early hepatoblasts or late hepatocytes are enriched for specific Gene Ontology biological processes. Heatmap  (using  Heatmap  Image  Creator  at  GenePattern:  www.broadinstitute.org/cancer/software/genepattern/) showing enrichment scores for the indicated Gene Ontology Biological Process terms (in rows) across the 14 gene clusters (in columns). Enrichment scores were calculated as the product of fold enrichment and the negative log10 of the respective Bonferroni value (Table 4-4). A low enrichment score indicates low-confidence enrichment of the respective GO term whereas a high score indicates high-confidence enrichment. Roman numerals at the top represent clusters with expression peak in the E10.5 (I), E12.5/E14.5 (II), E16.5 (III), or adult (IV) Tag-seq library. Note that cluster E exhibit constant expression and cluster I is bimodal, but are arbitrarily placed in the table for direct comparison with other clusters. GO enrichment analysis was performed using DAVID (http://david.abcc.ncifcrf.gov/summary.jsp; Huang et al., 2007). Gene symbols from each cluster were used as input dataset, with the 5449 genes included for K-means clustering used as background. 104  Chapter 5: Analysis of Hmga2 expression during hepatogenesis and in liver cell lines  5.1  Hmga2 is highly expressed among cluster A genes Amongst the 14 gene clusters generated from our clustering analysis, cluster A was  chosen for further analysis based on its well-defined temporal expression pattern that peaks in E10.5 DLK1+ cells (Figure 4-4). In addition, cluster A contains Smad2 and Yap1, two transcription factors that are known to control embryonic hepatoblast and post-natal hepatocyte growth (Dong et al., 2007; Septer et al., 2012; Weinstein et al., 2001). Moreover, cluster A genes also include Dnmt1, Lin28a, and Hmga2, which are transcription factors associated with stem cells and their self-renewal capacity (Li et al., 2006; Nishino et al., 2008; Sen et al., 2010; Trowbridge et al., 2009; Xu et al., 2009) but their roles have yet to be characterized during in vivo hepatoblast differentiation. Thus, further analysis of cluster A could identify new markers for early hepatoblasts and thus uncover hidden mechanisms that regulate hepatoblast differentiation. As a first step to analyzing cluster A genes, they were ranked according to their expression level in the E10.5 DLK1+ Tag-seq library (Table 5-1). The top three ranking genes are Hnrnpl, or heterogeneous nuclear ribonucleoprotein L, which regulates splicing efficiency (Cheli and Kunicki, 2006; Gaudreau et al., 2012; Liu and Mertz, 1995); Asb4, or ankyrin repeat and SOCS box-containing 4, which regulates vascular differentiation through ubiquitination and degradation of target proteins (Ferguson et al., 2007); and Hmga2, or high mobility group AT-hook 2, which regulates transcription by modulating chromatin structure (Ashar et al., 2010). To date, neither Hnrnpl nor Asb4 have been implicated in regulating 105  fetal or adult liver development. Interestingly, Hmga2 plays a role during liver regeneration in the mouse as it is activated in adult liver oval cells upon liver injury (Sakai et al., 2010). HMGA2 is highly expressed in human embryonic stem cells (Li et al., 2006), and Hmga2 has been implicated in lung bud formation in the mouse (Millien et al., 2008). The temporal expression pattern of Hmga2 as observed in the Tag-seq libraries was successfully validated by qRT-PCR (Figure 4-3). However, the role of HMGA2 during fetal hepatoblast specification and differentiation has yet to be explored. To begin examining this, I characterized the expression pattern of the HMGA2 protein in E9 and E9.5 liver buds, as well as E10.5, E12.5, and E14.5 livers.  5.1.1  HMGA2 is expressed in nascent hepatoblasts In E9.0 embryos, hepatic progenitors delaminate from the single-layered endoderm  and invade the septum transversum mesenchyme to form the liver bud. This is clearly shown in Figure 5-1A-D, where HNF4A-positive hepatic progenitors are encased by GATA4positive mesenchyme cells. Note the more intense HNF4A staining in the hepatic progenitors relative to other endoderm cells (compare staining intensity at thick dotted region with the thin dotted line in Figure 5-1B). When transverse E9.0 embryo sections were stained with an anti-HMGA2 antibody, the protein was observed in the mesenchyme as well as the entire endoderm monolayer (Figure 5-1E-H). In contrast to HNF4A, HMGA2 staining in the hepatic progenitors appeared less intense relative to non-hepatic endoderm cells (compare the thick dotted line against the thin in Figure 5-1F). At E9.5, the liver bud continues to expand into the septum transversum mesenchyme, and hepatoblasts retain strong expression of both HNF4A and HMGA2 (Figure 5-1J, N).  106  5.1.2  HMGA2 expression is retained in a subset of maturing hepatoblasts After showing HMGA2 to be expressed in nascent hepatoblasts, we next sought to  determine the proportion of hepatoblasts that express HMGA2, and specifically if this proportion changes as hepatoblasts mature during development. In the E9.5 liver bud, strong expression of DLK1 was already observed in hepatoblasts (Figure 5-2A-D). 69% (120/174) of E9.5 DLK1+ cells also stained for HMGA2 (Table 3-1), and the intensity of HMGA2 staining in DLK1-positive cells was similar to that of the DLK1-negative cells (Figure 5-2B, compare white arrows with yellow arrows). Interestingly, by E10.5, expression of HMGA2 became restricted to only 37.5% (78/208) of DLK1+ hepatoblasts (Figure 5-2E-H). Furthermore, HMGA2 staining in E10.5 DLK1+ cells became fainter as the intensity observed in DLK1-positive cells was significantly less compared to the DLK1-negative counterparts (Figure 5-2F, compare white against yellow arrows). By E12.5 and E14.5, HMGA2 expression was observed in only 12.2% (25/205) and 14.4% (35/243) of DLK1+ cells, respectively (Table 3-1). Similar to E10.5 fetal livers, HMGA2 staining intensity in E12.5 and E14.5 DLK-positive cells were lower relative to the DLK1-negative population (white arrows in Figure 5-3C and F). Taken together, HMGA2 is downregulated during hepatoblast maturation, resulting in a reduction in both the number of DLK1+ HMGA2+ fetal liver cells and the level of HMGA2 protein present in those cells.  5.2  Hmga2-/- embryos have fewer hepatoblasts Hmga2 has been proposed to maintain stem cells in an undifferentiated state (Ashar et  al., 2009; Li et al., 2006). The first study on Hmga2 mutant mice revealed a pygmy  107  phenotype, where homozygous mutant embryos exhibit growth retardation beginning at E15.5, culminating to adult mutants that are only 40% the size of wildtype (Benson and Chada, 1994). Hmga2-/- mutants exhibit a number of additional phenotypes, such as obesity and skeletal malformation (Moon et al., 2002), as well as a reduction in the number of neural stem cells (Nishino et al., 2008). However, defects in liver development and specifically hepatoblast differentiation have yet to be investigated in Hmga2 mutants. As HMGA2 is robustly expressed in nascent and less matured hepatoblasts, it may play a role in the expansion of hepatoblasts during development. To determine if Hmga2-/embryos have fewer hepatoblasts and thus smaller livers relative to their wildtype littermates, fetal livers with the respective phenotypes were collected and weighed. Minimal deviation in weight was observed between wildtype, Hmga2+/-, and Hmga2-/- E14.5 fetal livers (0.256g + 0.026g; 0.271g + 0.020g; and 0.268g + 0.019g, respectively; mean + SD). To more directly determine if Hmga2 mutants have fewer hepatoblasts, E14.5 livers from Hmga2-/- mutant were stained for DLK1+ cells and compared against wildtype and Hmga2 heterozygous littermates. DLK1+ fetal liver cells were observed at 4.37% + 1.00% (N=4) and 4.01% + 1.80% (N=16) in wildtype and Hmga2+/- embryos, respectively (mean + SD). However, in Hmga2-/- embryos, this figure is slightly decreased to 2.99% + 0.74% (mean + SD; N=6, p=0.04 in one-tailed T-test; Figure 5-4), implicating a role for Hmga2 in regulating hepatoblast number in the fetal liver. To determine if the reduced number of DLK1+ cells in Hmga2 mutants is due to the loss of DLK1 expression on hepatoblasts, E14.5 Hmga2 mutant liver sections were co-stained with antibodies against HNF4A and DLK1. As shown in Figure 5-5, HNF4A-positive hepatoblasts maintained strong expression of DLK1, and the proportion of HNF4A-positive cells co-expressing DLK1 was similar between Hmga2  108  heterozygous and homozygous mutants (819/826 or 99.1% versus 789/793 or 99.5%), suggesting DLK1 expression on hepatoblasts to be unaltered in the absence of HMGA2.  5.3  HMGA2 downregulates HNF4A in liver cell lines Several reports suggest Hmga2 to be inversely correlated with a hepatic phenotype.  For instance, fetal livers lacking functional HNF4A have under-developed hepatocytes and the expression of Hmga2 is upregulated (Battle et al., 2006; Parviz et al., 2003). Furthermore, hepatocyte-like cells generated from mouse fibroblasts show a dramatic upregulation of Hnf4! but also a large reduction of Hmga2 expression (Huang et al., 2011; Sekiya and Suzuki, 2011). In addition, Hnf4a has been shown to maintain the hepatic phenotype in hepatoma cells by downregulating Hmga2 and other mesenchymal genes (Santangelo et al., 2011). Lastly, through chromatin immunoprecipitation experiments in E14.5 DLK1+ cells, HNF4A was found to be associated with the Hmga2 locus (Figure 5-6), which further implies a role for HNF4A to regulate Hmga2. To further characterize the relationship between HNF4A and HMGA2 expression, HPPL cells were used. These hepatoblast-like, bipotential cells have been extensively characterized for their ability to differentiate down the hepatocyte lineage when grown to confluency under normal culturing conditions (Tanimizu et al., 2004). HPPLs were derived from MACS-isolated E14.5 DLK1+ cells as previously described (Tanimizu et al., 2003). When HPPLs were grown to confluency to induce hepatocyte differentiation, Hnf4! expression increased almost 10-fold relative to non-confluent cells (p=0.03; Figure 5-7A). Hmga2 is expressed in HPPL cells, but its expression level under high levels of Hnf4! has not been studied. To determine if Hnf4! regulates Hmga2, qRT-PCR was used to compare  109  Hmga2 level in HPPL cells grown to low confluency (when Hnf4! expression is low) versus high confluency (when Hnf4! expression is high). Hmga2 is readily detectable in low confluence HPPL cells, and although Hmga2 expression did not decrease in confluent HPPL cells, it showed an insignificant increase of less than 2-fold (p=0.13; Figure 5-7A). These data show that Hmga2 is not downregulated by Hnf4! in HPPL cells. Next, to determine if Hmga2 can downregulate Hnf4a, Hmga2 was ectopically expressed in HPPL cells. Specifically, lentivirus containing the full-length Hmga2 cDNA was used to transfect HPPL cells, and the transfected cells were assayed for Hnf4a expression at full confluency. Relative to the empty vector, HPPL cells transfected with Hmga2 showed a 2.5-fold increase in Hmga2 expression. In contrast, Hnf4a expression in the transfected cells was 24% lower than cells transfected with the empty vector control (p=0.03; Figure 5-7B), suggesting that Hmga2 weakly downregulates Hnf4a. To extend the analysis between Hnf4a and Hmga2 expression, I used human hepatocarcinoma HepG2 cells that simultaneously express robust levels of HNF4A and HMGA2 (Shell et al., 2007; Xie et al., 2009). These cells provide an ideal system for concurrent visualization of HNF4A and HMGA2 within the same cell. To do this, HepG2 cells co-stained with antibodies against HNF4A and HMGA2 were imaged using fluorescent microscopy, and a subset of cells with bright staining of either HNF4A or HMGA2 was selected for further analysis. For cells with bright HNF4A staining, 7 out of 8 cells stained dimly for HMGA2 (Figure 5-8A-H: compare HMGA2 staining in A with A’). On the other hand, 5 out of 8 HMGA2-bright cells showed dim staining for HNF4A (Figure 5-8A’-H’: compare HNF4A staining in A’ with A). These data suggest that high expression of HNF4A  110  and HMGA2 is mutually exclusive in HepG2 cells, and may further indicate Hmga2 to be a negative regulator of Hnf4!.  5.4 5.4.1  Discussion HMGA2 is dynamically expressed in a subset of fetal liver DLK1+ cells In this chapter, I presented the first report on the expression pattern of HMGA2  during in vivo hepatogenesis. HMGA2 was widely expressed in nascent hepatoblasts as the protein is found in close to 70% of E9.5 DLK1+ cells. HMGA2 is progressively downregulated as hepatoblasts mature over time, as the protein is observed in 37.5% and 14.4% of E10.5 and E14.5 DLK1+ cells, respectively (Table 3-1). As high expression of Hmga2 is associated with stem cell renewal (Li et al., 2006; Nishino et al., 2008), DLK1+ HMGA2+ cells could represent hepatoblasts with higher proliferation and differentiation potential. In a study by Tanaka et al., the epithelial cell adhesion molecule EpCAM is expressed in 24-34% of DLK1+ cells in the E12.5 and E14.5 fetal liver. Interestingly, DLK1+ EpCAM+ cells had a higher proliferation potential in vitro than DLK1+ EpCAM- counterparts, suggesting the double positive cells to be more progenitor-like (Tanaka et al., 2009). Thus, it would be of interest to determine if a subset of DLK1+ EpCAM+ cells also expresses HMGA2. However, the percentage of DLK1+ HMGA2+ in E14.5 fetal livers is comparable to that of DLK1+ HNF4A- cells, which are detected at 14.4% and 9.8%, respectively. Thus, triple labeling of fetal liver sections with antibodies against DLK1, HNF4A, and HMGA2 could be performed to ensure that DLK1+ HMGA2+ cells in the E14.5 fetal liver are still hepatic. This experiment was not performed as the primary antibodies used in this study to detect HNF4A and HMGA2 were both raised in rabbits, thereby prohibiting unambiguous  111  detection of the proteins by the secondary antibody. Primary antibodies raised in different species or directly conjugated to a fluorochrome would be required for the triple-labeling assay. Data from Tag-seq libraries and antibody staining experiments in this study suggest discordance between the transcript and protein levels of Hmga2. For instance, at E10.5, HMGA2 protein is expressed in 38% of DLK1+ cells (Table 3-1), and Hmga2 transcripts are detected in the E10.5 DLK1+ Tag-seq library at 520 TPM. In comparison, 90% of DLK1positive cells expressed the hepatoblast marker HNF4A, and Hnf4a is expressed in E10.5 Tag-seq library at 85 TPM. Thus, while Tag-seq detected Hmga2 mRNA at a higher level than Hnf4a, the percentage of E10.5 DLK1+ cells expressing the HMGA2 protein is lower than expected. This observation could be explained by high Hmga2 transcriptional activity being restricted to a small subset of DLK1+ cells, and the experiment described above investigating HMGA2 expression in the subset of DLK1+ EpCAM+ fetal liver cells could directly test this. Alternatively, Hmga2 expression could be regulated post-transcriptionally. Specifically, miRNAs could bind imperfectly (ie. with base pair mismatches) to the 3’ UTR of the Hmga2 mRNA to inhibit mRNA translation (Fabian et al., 2010). miRNAs have been shown to be dynamically expressed during fetal liver development (Liu et al., 2010), and in vitro experiments have confirmed a role for miRNAs in the miR-23b cluster to regulate hepatoblast differentiation (Rogler et al., 2009). Thus, identifying miRNAs that can bind to the Hmga2 mRNA 3’ UTR and are also drastically upregulated in hepatoblasts between E9.5 and E10.5 would help determine if miRNAs play a role in controlling the development of midgestation hepatoblasts.  112  5.4.2  HMGA2 may regulate two distinct phases of liver development A number of transcription factors with dual roles in hepatogenesis have been  identified, such as HNF6 (Clotman et al., 2002; Margagliotti et al., 2007), TBX3 (Ludtke et al., 2009; Suzuki et al., 2008), HEX (Bort et al., 2006; Hunter et al., 2007), and FOXM1B (Krupczak-Hollis et al., 2004). In this chapter, I presented data that Hmga2-/- E14.5 embryos have fewer hepatoblasts, and Hmga2 can downregulate Hnf4! in liver cell lines. These findings suggest that HMGA2 may be a transcription factor that has roles in both the proliferation and differentiation phases of hepatoblast maturation. However, several experimental factors and limitations surrounding these findings must first be addressed before solidifying such a claim. Fetal livers from E14.5 Hmga2 null embryos were found to have fewer hepatoblasts relative to wildtype (Figure 5-4). This strongly implies that Hmga2 has a role in generating hepatoblasts in the developing fetal liver. However, as the sample size for this analysis is low, continuing the analysis with more Hmga2 mutant littermates would increase the statistical certainty of this claim. In addition, the observed reduction of hepatoblast number in Hmga2-/- embryos is only 30% relative to wildtype. This suggests that either Hmga2 is not necessary for hepatoblasts proliferation, or functional redundancy with Hmga1 and other HMG proteins are compensating for the loss of Hmga2 (Beitzel and Bushman, 2003). Western blot experiments comparing the expression levels of HMG proteins in Hmga2-/versus wildtype E14.5 DLK1+ cells would help determine if functional redundancy underlies the slight reduction in hepatoblast number. Moreover, in the fetal liver, HMGA2 is expressed in both hepatic and non-hepatic cells (Figure 5-2 and Figure 5-3). As the Hmga2-/- mutants used in this study are non-conditional knockouts, the observed phenotype could be caused by  113  the lack of functional Hmga2 in hepatoblasts, non-hepatoblasts, or both cell types. Generating conditional Hmga2 knockouts using mouse lines with Cre recombinase expression driven by either a hepatoblast or mesenchyme-specific promoter can shed light into the cell type(s) that is responsible for regulating hepatoblast number in the fetal liver. In vitro experiments using the HPPL and HepG2 liver cell lines generated two lines of evidence for Hmga2 to weakly downregulate Hnf4!. First, HPPL cells with a 2.5-fold ectopic expression of Hmga2 showed a reduction in Hnf4! mRNA level (Figure 5-7B). While the decrease was only 24%, the high reproducibility of this reduction (a mean value of 0.76 for a standard deviation of 0.018) increases the validity of this claim. However, as the pvalue of the reduction is just two percentage points below the 0.05 threshold, performing more biological replicates should increase the statistical significance of these findings. Secondly, using HepG2 human hepatoma cells, 5 out of 8 cells with high levels of HMGA2 had low expression of HNF4A (Figure 5-8). While this finding suggests an inverse relationship between the protein levels of HMGA2 and HNF4A, the correlation is not absolute as three HMGA2-bright cells expressed HNF4A at high levels. Furthermore, as this comparison was performed by a human eye and limited to HepG2 cells expressing high levels of HMGA2, the analysis yielded qualitative data on a small and specific subset of cells. Analyzing antibody-stained HepG2 cells using dual-colour intracellular FACS experiments (Festuccia and Chambers, 2011; Lafarge et al., 2007) would provide a more global, objective, and quantitative comparison of HNF4A and HMGA2 levels. Moving forward, it would be of interest to determine if HMGA2-mediated downregulation of HNF4A can be extended into the in vivo context of developing hepatoblasts. Hmga2 mutant embryos have fewer hepatoblasts, and these hepatoblasts retain robust expression of HNF4A (Figure  114  5-5). However, as this data was obtained qualitatively via immunofluorescence microscopy, it is insufficient for determining if HNF4A levels in Hmga2-/- hepatoblasts are increased relative to wildtype littermates. Therefore, further Western blot or qRT-PCR experiments using Hmga2-/- and wildtype hepatoblasts are needed to determine if Hnf4! expression is regulated by HMGA2 during in vivo hepatoblast maturation. A previous study has shown HNF4A to transcriptionally repress HMGA2 in Hep3B human hepatoma and BW1J mouse hepatoma cells (Santangelo et al., 2011). In this chapter, I showed that high protein levels of HNF4A are correlated with low HMGA2 expression in HepG2 cells (Figure 5-8). Although this analysis was limited to 8 HNF4A-bright cells, all 8 cells stained dimly for HMGA2, which provides preliminary data to support a role for HNF4A to downregulate HMGA2. However, as stated above, such analysis is not quantitative and prone to observational bias. Therefore, analyzing the HepG2 cells using dual-colour intracellular FACS would provide data that is more objective and quantitative. In contrast to HepG2 cells, Hnf4! and Hmga2 expression was not inversely correlated in HPPL cells. Here, when Hnf4! expression was elevated in HPPL cells by culturing to full confluency, no reduction in Hmga2 level was observed (Figure 5-7A). One possibility is that the increase of Hnf4! expression was insufficient for causing repression of Hmga2, and thus other methods to ectopically express Hnf4! in HPPL cells may be required before an effect on Hmga2 expression is observed. Alternatively, Hnf4! may downregulate Hmga2 in a context-dependent manner. HPPL cells are derived from E14.5 fetal liver DLK1+ cells (Tanimizu et al., 2003), whereas HepG2, Hep3B, and BW1J are all hepatoma cell lines (Aden et al., 1979; Knowles et al., 1980; Szpirer and Szpirer, 1975). Thus, the ability of  115  HNF4A to downregulate HMGA2 may only be observed when HNF4A is expressed above normal physiological levels, such as in cancers. In conclusion, further experiments are needed in order to confirm the possible crosstalk between Hmga2 and Hnf4! during liver development. Specifically, experiments using the Hmga2 knockout strain will provide in vivo data in determining if Hmga2 regulates Hnf4! and potentially other hepatoblast differentiation genes. HMGA2 controls gene expression by modulating local chromatin structure (Cleynen and Van de Ven, 2008), and this mechanism of gene regulation could allow HMGA2 to act on multiple hepatoblast differentiation genes. Genomic regions bound by HMGA2 have been mapped in a colon cancer cell line via chromatin precipitation (Winter et al., 2011). Although the study did not list HNF4A as one of the HMGA2-bound loci, its data can nonetheless be mined for additional transcriptional HMGA2 targets, some of which could be genes involved in hepatoblast differentiation. In addition, knockout mice of transcription factors that regulate hepatoblast differentiation have been generated, and their phenotypes during in vivo hepatoblast maturation have been thoroughly described (Clotman et al., 2002; Hunter et al., 2007; Margagliotti et al., 2007; Ludtke et al., 2009; Suzuki et al., 2008). Via in situ hybridization and qRT-PCR experiments, the expression of hepatocyte (such as Cebp! and Hnf4!) and cholangiocyte (such as Onecut1, Onecut2, Hnf1", and Hnf6) markers were found to be misregulated in hepatoblasts of the respective mutants, resulting in the formation of hybrid hepatoblast-like cells that express both hepatocyte and biliary markers even in late gestation livers. Similar experiments can be performed in Hmga2 null embryos to determine if hepatocyte and cholangiocyte markers are also misregulated during development of the mutant embryonic livers.  116  On the other hand, to determine if Hnf4! can downregulate Hmga2 during normal hepatogenesis, hepatoblasts from fetal liver-specific Hnf4! knockouts can be assayed for increased expression of Hmga2. The molecular mechanism underlying HNF4A-mediated suppression of gene expression has been examined (Lazarevich, 2000). In E14.5 DLK1+ cells, HNF4A occupies two intronic regions of the Hmga2, implicating HNF4A to have a regulatory role on Hmga2 expression during hepatoblast development in vivo (Figure 5-6). ChIP-seq and ChIP-chip studies have identified upwards of 10,000 genomic sites bound by HNF4A during mammalian organogenesis, and intronic peaks were proposed to still be functionally relevant in regulating gene expression (Hoffman et al., 2010; Kyrmizi et al., 2006; Odom et al., 2006; O. Alder, personal communication). Therefore, further experiments to elucidate the functional significance of the HNF4A peaks in the Hmga2 locus would clarify the potential crosstalk that between Hnf4! and Hmga2 during hepatoblast maturation in the mouse embryo. For instance, luciferase assays using in HPPL cells or other fetal hepatoblast cell lines such as HBC-3 (Rogler, 1997) can be performed to determine if the HNF4A-bound regions can negatively impact Hmga2 expression in the context of fetal hepatoblasts.  117  Gene  Description  1  TPM in E10.5 DLK1+ Tag-seq library 1268  Hnrnpl  2  549  Asb4  3 4 5 6 7 8 9 10  520 438 391 323 311 262 261 252  Hmga2 Rps26l Marcksl1 Nap1l1 Sfrs7 Prmt5 Ranbp1 Snrpb2  heterogeneous nuclear ribonucleoprotein L ankyrin repeat and SOCS boxcontaining 4 high mobility group AT-hook 2 ribosomal protein S27-like MARCKS-like 1 nucleosome assembly protein 1-like 1 serine/arginine-rich splicing factor 7 protein arginine N-methyltransferase 5 RAN binding protein 1 U2 small nuclear ribonucleoprotein B  Rank  Table 5-1 Top 10 genes in cluster A as ranked by expression (in TPM) in the E10.5 DLK1+ Tag-seq library  118  119  Figure 5-1 HMGA2 is expressed in E9 and E9.5 hepatoblasts. (A-D) Transverse section through an E9.0 embryo to was stained with anti-HNF4A and anti-GATA4 antibodies to visualize the newly formed liver bud and nascent hepatoblasts. At low magnification (A), the liver bud can be recognized by its characteristic bulb-like structure protruding into the septum transversum mesenchyme. Higher magnification shows nascent hepatoblasts with intense HNF4A staining (B, thick dotted line) but absent for GATA4 (C). The image in B is over-exposed to highlight hepatoblasts and in the process makes non-hepatoblasts faintly positive. (E-H) Embryo section in the same orientation but stained for HMGA2 and GATA4. HMGA2 is present in the monolayer endoderm, including hepatoblasts (E). Higher magnification of a serial section shows HMGA2 staining in nascent hepatoblasts to be less intense relative to other endodermal cells (F, compare thick dotted line to thin dotted line). HMGA2 is present in both the hepatoblasts and the neighbouring septum transversum mesenchyme, as depicted by HMGA2 and GATA4 double positive cells that appear yellow in panel H. (I-L) Transverse section through an E9.5 embryo was stained with anti-HNF4A and antiGATA4 antibodies to visualize hepatoblast invasion into the septum transversum mesenchyme. As in E9.0, hepatoblasts stain brightly for HNF4A (J) but absent for GATA4 (K). (M-P) Embryo section in the same orientation but stained for HMGA2 and GATA4. Similar to E9.0, HMGA2 is expressed in hepatoblasts and the septum transversum mesenchyme, shown as yellow cells in panel P. Boxed areas in panels A, E, I, and M indicate enlarged regions, shown to the right. Scale bar respectively represents 100um, 50um, and 25um at 100x, 200x, and 400x magnification.  120  Figure 5-2 A subset of DLK1+ hepatoblasts also express HMGA2 in the E9.5 liver bud and E10.5 liver. Deconvolved immunofluorescent images of E9.5 and E10.5 fetal liver sections stained with anti-DLK1 and anti-HMGA2 antibodies. (A-D) At E9.5, DLK1+ hepatoblasts expressing HMGA2 were detected (white arrows), and staining intensity of HMGA2 is comparable to that of DLK1- cells (yellow arrows). (E-H) At E10.5, HMGA2 was detected in some hepatoblasts (white arrows). However, HMGA2 staining is less intense compared to non-hepatoblasts (yellow arrows). Scale bar = 100!m at 100x, 50!m at 200x, and 25!m at 400x magnification.  121  Figure 5-3 Few DLK1+ cells co-express HMGA2 in E12.5 and E14.5 fetal liver. Deconvolved immunofluorescent images of E12.5 and E14.5 fetal liver sections stained with anti-DLK1 and anti-HMGA2 antibodies. At both stages, HMGA2 is expressed in non-hepatic cells, and HMGA2-positive hepatoblasts were scarcely detected (white arrows). Scale bar = 25!m.  122  Figure 5-4 Fewer fetal liver DLK1+ cells are present in Hmga2-/- embryos. E14.5 fetal livers cells from the indicated genotypes were collected, stained with an antiDLK1 antibody, and analyzed by FACS. The percentage of DLK1+ fetal liver cells was reduced in Hmga2 homozygotes relative to heterozygous littermates (p=0.04 in a one-tailed T-test).  123  Figure 5-5 Hmga2 mutant fetal hepatoblasts retain DLK1 expression. Immunofluorescent images of E14.5 sectioned fetal livers from three Hmga2 heterozygotes (A to C) or three homozygotes (D-F) costained with antibodies to detect DLK1 (green) and HNF4A (red). DAPI-positive cell nuclei are false-coloured blue. Scale bar = 50!m.  124  Figure 5-6 HNF4A occupies several sites within Hmga2 in E14.5 DLK1+ hepatoblasts. E14.5 fetal liver DLK1+ cells were isolated and subjected to chromatin immunoprecipitation experiments using a ChIP-grade antiHNF4A antibody. The precipitated chromatin was sequenced, aligned to the mouse genome, and peaks around the Hmga2 locus are shown here in a UCSC browser screenshot. Direction of the Hmga2 gene track goes from right to left. The first HNF4A peak is located 40kb upstream of transcriptional start site (TSS); both the second and third peaks are found in the third intron, with the second peak being 60kb downstream from the TSS, and the third peak a further 27kb downstream. Numbers on the left denote peak height, which is indicative of binding strength.  125  Figure 5-7 Hmga2 expression in HPPL cells. (A) HPPL cells were grown to low (20-30%), medium (40-60%), or high (80-90%) confluency before total RNA was harvested for quantitative RT-PCR. Expression of hepatic markers Afp and Hnf4! increased at higher confluency, while confluency had minimal effects on Hmga2. (B) HPPL cells were infected with either the MIY or Hmga2-MIY expression construct, replated once to expand cells to full confluency, and confluent cells collected for RNA extraction and quantitative RT-PCR. Ectopic Hmga2 expression led to a statistically significant reduction in the level of Hnf4a but not Afp.  126  Figure 5-8 Expression level of HMGA2 and HNF4A are inversely correlated in HepG2 cells. Immunofluorescent images of HepG2 cells stained with anti-HNF4A and anti-HMGA2 antibodies. Cells with the same alphabetical designation (for instance, A and A’) were imaged taken in the same field of view and equal exposure times to allow direct comparison of protein expression levels. In 8 select cells with bright staining of HNF4A, HMGA2 staining was dim in 7 cases (panels A-C, E-H). On the other hand, in 8 select cells with bright HMGA2 staining, 5 out of 8 cases (panels A, D, E, G, and H) showed dim HNF4A staining. Scale bar = 10!m. 127  Chapter 6: Future directions  6.1  Thesis overview In this thesis, the expression of the cell surface molecule DLK1 was characterized in  the mouse fetal liver. The protein was specifically enriched on hepatoblasts but not other fetal liver cell types. Fetal liver DLK1+ cells were isolated at E10.5, E12.5, E14.5, and E16.5, and their respective transcriptomes were captured in Tag-seq libraries. The resulting data were highly enriched for fetal hepatic genes, suggesting the libraries to be an accurate portrayal of the transcriptome in maturing hepatoblasts. Subjecting the Tag-seq data to Kmeans clustering analysis revealed 14 gene clusters with distinct patterns of temporal gene expression. Gene Ontology analysis showed that many transcription factors are highly expressed in E10.5 hepatoblasts but are drastically reduced in hepatoblasts of subsequent stages. Further analysis focused on Hmga2, a gene associated with stem cell maintenance but has no reported role during liver development. E14.5 Hmga2-/- embryos showed a slight reduction in hepatoblast number, and high expression of HMGA2 in HPPL and HepG2 cells suppressed expression of the hepatic marker HNF4A. These results suggest Hmga2 to potentially regulate both hepatoblast proliferation and differentiation during hepatogenesis.  6.2  Deep sequencing libraries from fetal liver DLK1+ cells will shed light into the  defining characteristic of fetal hepatoblasts To date, my in-depth analysis showing DLK1-positive cells to be enriched for fetal hepatoblasts has provided a strong basis for studying hepatoblast differentiation. In addition to using fetal liver DLK1+ cells for constructing Tag-seq libraries described here, our lab has  128  also used the E14.5 DLK1+ population to generate additional next-generation sequencing libraries to address other biological questions, such as miRNA expression (Wei et al., in press), nucleosome positioning (Hoffman et al., 2010), and transcription factor binding (O. Alder, personal communication). Specifically, by overlaying genome-wide transcription factor occupancy data with gene expression data, our lab has recently discovered that HNF4A mediates developmental stage-specific gene expression by preferential recruitment of cofactors (Alder et al., manuscript submitted). Therefore, a comprehensive analysis combining the transcriptome, miRNA-ome, and interactome data generated using E14.5 DLK1+ hepatoblasts should yield unprecedented insights into the genetic and epigenetic switches in fetal hepatoblasts. The DLK1 antigen (Tanimizu et al., 2003) was used to isolate hepatoblasts during embryogenesis and construct Tag-seq libraries. DLK1+ cells have been isolated for detailed in vitro and in vivo studies. Mouse fetal liver E14.5 DLK1+ cells can home into an injured adult liver and give rise to new hepatocytes during liver regeneration (Tanimizu et al., 2003). Furthermore, primary DLK1+ cells were used to study the effects of Notch signaling on hepatoblast differentiation (Tanimizu and Miyajima, 2004), and HPPL cells derived from primary DLK1+ cells have been used as a system for identifying growth conditions needed to promote hepatocyte or cholangiocyte lineage commitment (Tanimizu et al., 2004; Tanimizu et al., 2007). Interestingly, the function of DLK1 is conserved in rat fetal livers, as DLK is specifically expressed on fetal liver hepatoblasts and E14 DLK1+ hepatoblasts could repopulate the adult liver after partial hepatectomy (Oertel et al., 2008). Thus, collectively, the Tag-seq libraries in this study represent a novel resource for the transcriptome of DLK1expressing cells. More in-depth analysis of this resource could identify genes and genetic  129  mechanisms that not only underlie hepatoblast differentiation but also the ability for hepatoblasts to participate in liver regeneration.  6.3  Tag-seq libraries will serve as a useful resource for identifying markers of liver  stem and progenitor cells Analysis of the Tag-seq libraries generated in this study unraveled dynamic gene expression signatures during hepatoblast maturation. As this was the initial study using nextgeneration sequencing to analyze gene expression changes during in vivo hepatoblast differentiation, library analysis was limited to known transcripts. Future efforts could concentrate on analyzing other RNA species such as anti-sense and non-protein coding RNA. These RNA species have been shown to be important regulators in various aspects of mouse development (Bevilacqua et al., 1988; Rapicavoli et al., 2010; Rosenbluh et al., 2011; Sekine et al., 2009; Yang et al., 2005), and detection algorithms for these non-protein coding transcripts in next-generation sequencing transcriptome libraries have been described (Fasold et al., 2011; Morrissy, 2010). Transcriptome data from other fetal and adult liver cells are already available. Thus, cross-comparison of the DLK1+ Tag-seq libraries against gene expression datasets from other liver cells can shed light into different aspects of liver biology. For instance, hematopoietic and hepatic cells in the fetal liver reciprocally support expansion of the other cell type via the secretion of soluble cytokines such as oncostatin M (Kamiya et al., 1999), angiopoietin-like 3, and IGF2 (Chou and Lodish, 2010). Thus, comparing our transcriptome data against that of whole fetal livers (Jochheim-Richter et al., 2006; Li et al, 2008; Otu et al., 2007) may identify potential ligand-receptor relationships underlying the observed reciprocal growth  130  control. Also, transcriptome data is available for hepatocyte-like cells induced from adult fibroblasts (Huang et al., 2011; Sekiya and Suzuki, 2011). Interestingly, these hepatocyte-like cells activate some hepatoblast genes such as Afp and CK19 (Huang et al., 2011). Thus, identifying highly expressed genes that are common to both DLK1+ hepatoblasts and hepatocyte-like cells may indicate genes that need to be repressed in order to obtain fully functional hepatocytes. Finally, methods for isolating progenitor cells from the adult liver have been recently described, and their transcriptomes have been captured with microarrays (Dorrell et al., 2011; Furuyama et al., 2011; Shin et al., 2011). Similar to fetal liver DLK1+ cells, these adult liver progenitor cells can differentiate into hepatocytes and cholangiocytes in vitro, as well repopulate an injured adult liver. Thus, identifying common features between the transcriptomes of DLK1+ fetal hepatoblasts and adult liver stem cells (Dorrell et al., 2011; Shin et al., 2011) may lead to the identification of master regulators that are central to all hepatic progenitors. Through K-means clustering, I identified gene clusters with distinct expression patterns during the course of hepatoblast maturation. Significantly, some clusters show peak expression at E10.5, while others showed expression maxima at the E12.5/14.5 or E16.5 timepoints. These clusters could contain stage-specific genes that can be used as markers for gauging the in vitro hepatocyte-generating protocols (Soto-Gutierrez et al., 2007; Roelandt et al., 2010). Furthermore, as the respective clusters also contain genes that code for cytokines and extracellular matrix molecules, they shed light into potential autocrine and paracrine signaling that can be used to optimize in vitro differentiation of hepatocytes. Lastly, as hepatocytes continue to mature after birth (Kanamura et al., 1985), combining our Tag-seq libraries with postnatal hepatocytes transcriptome data (Li et al., 2009) can provide the full  131  scope of sequential gene expression changes that accompany the generation of terminally differentiated hepatocytes.  6.4  Applications of Tag-seq libraries toward improving current methods of liver  disease treatments In addition to using the Tag-seq libraries to characterize hepatic progenitors, the transcriptome data can also be used to gain insights into other liver-related processes. Alphafetoprotein, a fetal liver-specific molecule, has long been one of the most commonly used diagnostic markers for hepatocellular carcinoma, or HCC (Yuen and Lai, 2005). Furthermore, pro-mitotic genes expressed in fetal hepatocytes are often reactivated in human liver cancer samples (Coulouarn et al., 2005). Thus, as HCC often reactivates genes that are expressed in fetal hepatocytes, the transcriptome data in this thesis can be compared against that of HCC patients to determine how much of the liver cancer transcriptome is recapitulating developmental programs. Historically, histological information such as nodule size and tumour invasiveness have been used to categorize liver cancer biopsies into different grades and stages (Pons et al., 2005), and this information is interpreted by clinicians to determine the most suitable type of treatment for patients (Marrero et al., 2010). More recently, different grades of HCC have been associated with distinct gene expression signatures (Lee et al., 2004; Maass et al., 2010; Pons et al., 2005). Thus, it would be of interest to compare the transcriptome dynamics during hepatogenesis and liver carcinogenesis. This comparison may identify developmental pathways that are aberrantly activated during liver cancer, and therapeutic agents inhibiting these pathways could be used to revert tumourigenic liver cells to their regular state.  132  Liver regeneration is another biological phenomenon with similarities to the embryonic liver. Hepatoblasts markers DLK1, GPC3, and EPCAM have all been shown to be activated in either oval cells or hepatic stellate cells during regeneration (do Boer et al., 1999; Grozdanov et al., 2006; Tanaka et al., 2009; Tanimizu et al., 2003; Okabe et al., 2009; Zhu et al., 2012). This implies that liver regeneration reactivates genes that are important during fetal liver development. When the adult liver regenerates, quiescent hepatocytes undergo up to two rounds of mitosis by re-activating cell cycle genes that are expressed during expansion of fetal hepatoblasts (Fausto, 2000; Otu et al., 2007). The regeneration process also recruits inflammatory cells to the injury site and activates hepatic stellate cells, which aid tissue repair by depositing the extracellular matrix molecules necessary for tissue remodeling (Sano et al., 2010). However, prolonged deposition of extracellular cellular matrix molecules results in tissue scarring, commonly known as liver fibrosis (Ijzer, 2008). Interestingly, DLK1 is specifically reactivated in hepatic stellate cells during liver regeneration and fibrosis in the adult rat liver (Zhu et al., 2012). Data in this thesis suggest that molecules involved in wound healing and inflammatory responses are readily expressed in late gestation DLK1+ fetal liver cells (Figure 4-5, Table 4-4). Therefore, deeper understanding of fetal liver DLK1+ cells may uncover new mechanism underlying a controlled process of tissue remodeling, and this knowledge can be applied to reduce the occurrence of fibrosis in post-regeneration livers. Lastly, Tag-seq libraries in this study could prove to be an invaluable resource for optimizing current methods of in vitro hepatocyte generation. Current protocols suffer from inefficient and incomplete differentiation, which could be due to culturing conditions that do no fully mimic the genetic programs that are activated during in utero hepatocyte differentiation (Baxter et al., 2010; Huang et al., 2011; Sekiya and Suzuki, 2011; Soto-  133  Gutierrez et al., 2007; Roelandt et al., 2010). The temporal transcriptome data described in this study will be instrumental in deciphering developmental programs that are active during in utero hepatogenesis. Specifically, they can be mined for stage-specific markers during hepatoblast differentiation, and these markers can be used to assess the differentiation state of hepatocyte-like cells. By identifying genes that are differentially expressed during the process of inducing human pluripotent stem cells into hepatocyte-like cells, a set of temporal differentiation markers was identified (DeLaForest et al., 2011). Moreover, these markers were successful in characterizing hepatocytes generated using patient-derived stem cells (Cayo et al., 2012). These studies highlight the need for stage-specific markers as hepatoblasts differentiate in utero, as these markers will better reflect the genetic programs activated during fetal hepatogenesis and thereby assist the pursuit for generating truly functional hepatocytes under laboratory conditions. These hepatocytes will have a wide range of clinical implications. For instance, they can be expanded in culture and used as in vitro models for testing drug toxicity, a procedure that is currently limited to using primary hepatocytes (Baxter et al., 2010; Vanhaecke and Rogiers, 2006). Furthermore, they can serve as the raw biological material to be placed onto scaffolds to generate bioartificial livers (Strain and Neuberger, 2002; Vosough et al., 2011), which in turn will greatly alleviate the current demand for donor organs to treat end-stage liver disease patients. Lastly, the fully differentiated hepatocytes generated in vitro can be used as a controlled supply of cells for optimizing liver-targeted gene therapy, a procedure for treating heritable metabolic liver diseases such as familial hypercholesterolemia and ornithine transcarbamylase deficiency (Ghosh et al., 2000). Overall, the potential to generate fully functional, clinical-grade hepatocytes hold great promise as a sustainable method to  134  treating patients with liver diseases. In conclusion, as technological advances continue to reduce the cost of sequencing, transcriptome profiling will become increasingly relevant in clinical settings (Morgan et al, 2012). Future research efforts to create effective and userfriendly analysis pipelines will allow maximal utility of genome-wide sequencing studies.  135  References  Abelev, G. I. (1968). "Production of embryonal serum alpha-globulin by hepatomas: review of experimental and clinical data." Cancer Res 28(7): 1344-50. Aden, D. P., A. Fogel, et al. (1979). "Controlled synthesis of HBsAg in a differentiated human liver carcinoma-derived cell line." Nature 282(5739): 615-6. Ader, T., R. Norel, et al. (2006). "Transcriptional profiling implicates TGFbeta/BMP and Notch signaling pathways in ductular differentiation of fetal murine hepatoblasts." Mech Dev 123(2): 177-94. Akagi, K., R. M. Stephens, et al. (2010). "MouseIndelDB: a database integrating genomic indel polymorphisms that distinguish mouse strains." Nucleic Acids Res 38(Database issue): D600-6. Allocco, D. J., I. S. Kohane, et al. (2004). "Quantifying the relationship between co-expression, coregulation and gene function." BMC Bioinformatics 5: 18. Anisimov, S. V., K. V. Tarasov, et al. (2002). "SAGE identification of gene transcripts with profiles unique to pluripotent mouse R1 embryonic stem cells." Genomics 79(2): 169-76. Antoniou, A., P. Raynaud, et al. (2009). "Intrahepatic bile ducts develop according to a new mode of tubulogenesis regulated by the transcription factor SOX9." Gastroenterology 136(7): 232533. Ashar, H. R., R. A. Chouinard, Jr., et al. (2010). "In vivo modulation of HMGA2 expression." Biochim Biophys Acta 1799(1-2): 55-61. Ashburner, M., C. A. Ball, et al. (2000). "Gene ontology: tool for the unification of biology. The Gene Ontology Consortium." Nat Genet 25(1): 25-9. Audic, S. and J. M. Claverie (1997). "The significance of digital gene expression profiles." Genome Res 7(10): 986-95. Barrett, M. T., J. Glogovac, et al. (2002). "High-quality RNA and DNA from flow cytometrically sorted human epithelial cells and tissues." Biotechniques 32(4): 888-90, 892, 894, 896. Battle, M. A., G. Konopka, et al. (2006). "Hepatocyte nuclear factor 4alpha orchestrates expression of cell adhesion proteins during the epithelial transformation of the developing liver." Proc Natl Acad Sci U S A 103(22): 8419-24. Baxter, M. A., C. Rowe, et al. (2010). "Generating hepatic cell lineages from pluripotent stem cells for drug toxicity screening." Stem Cell Res 5(1): 4-22. Beitzel, B. and F. Bushman (2003). "Construction and analysis of cells lacking the HMGA gene 136  family." Nucleic Acids Res 31(17): 5025-32. Benson, K. F. and K. Chada (1994). "Mini-mouse: phenotypic characterization of a transgenic insertional mutant allelic to pygmy." Genet Res 64(1): 27-33. Bermingham-McDonogh, O., E. C. Oesterle, et al. (2006). "Expression of Prox1 during mouse cochlear development." J Comp Neurol 496(2): 172-86. Bevilacqua, A., R. P. Erickson, et al. (1988). "Antisense RNA inhibits endogenous gene expression in mouse preimplantation embryos: lack of double-stranded RNA "melting" activity." Proc Natl Acad Sci U S A 85(3): 831-5. Birney, E., T. D. Andrews, et al. (2004). "An overview of Ensembl." Genome Res 14(5): 925-8. Blackshaw, S., S. Harpavat, et al. (2004). "Genomic analysis of mouse retinal development." PLoS Biol 2(9): E247. Bochkis, I. M., N. E. Rubins, et al. (2008). "Hepatocyte-specific ablation of Foxa2 alters bile acid homeostasis and results in endoplasmic reticulum stress." Nat Med 14(8): 828-36. Bonzo, J. A., C. H. Ferry, et al. (2012). "Suppression of hepatocyte proliferation by hepatocyte nuclear factor 4alpha in adult mice." J Biol Chem 287(10): 7345-56. Bort, R., M. Signore, et al. (2006). "Hex homeobox gene controls the transition of the endoderm to a pseudostratified, cell emergent epithelium for liver bud development." Dev Biol 290(1): 4456. Bossard, P. and K. S. Zaret (1998). "GATA transcription factors as potentiators of gut endoderm differentiation." Development 125(24): 4909-17. Bouzin, C., F. Clotman, et al. (2003). "The onecut transcription factor hepatocyte nuclear factor-6 controls B lymphopoiesis in fetal liver." J Immunol 171(3): 1297-303. Bowie, M. B., K. D. McKnight, et al. (2006). "Hematopoietic stem cells proliferate until after birth and show a reversible phase-specific engraftment defect." J Clin Invest 116(10): 2808-16. Bray, S. J., S. Takada, et al. (2008). "The atypical mammalian ligand Delta-like homologue 1 (Dlk1) can regulate Notch signalling in Drosophila." BMC Dev Biol 8: 11. Butte, A. (2002). "The use and analysis of microarray data." Nat Rev Drug Discov 1(12): 951-60. Cai, L., H. Huang, et al. (2004). "Clustering analysis of SAGE data using a Poisson approach." Genome Biol 5(7): R51. Cantz, T., D. M. Zuckerman, et al. (2003). "Quantitative gene expression analysis reveals transition of fetal liver progenitor cells to mature hepatocytes after transplantation in uPA/RAG-2 mice." Am J Pathol 162(1): 37-45.  137  Carlsson, C., D. Tornehave, et al. (1997). "Growth hormone and prolactin stimulate the expression of rat preadipocyte factor-1/delta-like protein in pancreatic islets: molecular cloning and expression pattern during development and growth of the endocrine pancreas." Endocrinology 138(9): 3940-8. Cascio, S. and K. S. Zaret (1991). "Hepatocyte differentiation initiates during endodermalmesenchymal interactions prior to liver formation." Development 113(1): 217-25. Cayo, M. A., J. Cai, et al. (2012). "JD induced pluripotent stem cell-derived hepatocytes faithfully recapitulate the pathophysiology of familial hypercholesterolemia." Hepatology 56(6): 216371. Challita, P. M., D. Skelton, et al. (1995). "Multiple modifications in cis elements of the long terminal repeat of retroviral vectors lead to increased expression and decreased DNA methylation in embryonic carcinoma cells." J Virol 69(2): 748-55. Cheli, Y. and T. J. Kunicki (2006). "hnRNP L regulates differences in expression of mouse integrin alpha2beta1." Blood 107(11): 4391-8. Chen, J., M. Sun, et al. (2002). "Identifying novel transcripts and novel genes in the human genome by using novel SAGE tags." Proc Natl Acad Sci U S A 99(19): 12257-62. Cherqui, S., S. M. Kurian, et al. (2006). "Isolation and angiogenesis by endothelial progenitors in the fetal liver." Stem Cells 24(1): 44-54. Chou, S. and H. F. Lodish (2010). "Fetal liver hepatic progenitors are supportive stromal cells for hematopoietic stem cells." Proc Natl Acad Sci U S A 107(17): 7799-804. Cleynen, I. and W. J. Van de Ven (2008). "The HMGA proteins: a myriad of functions (Review)." Int J Oncol 32(2): 289-305. Clissold, P. M. and J. O. Bishop (1981). "Molecular cloning of cDNA sequences transcribed from mouse liver endoplasmic reticulum poly(A)mRNA." Gene 15(2-3): 225-35. Clotman, F., P. Jacquemin, et al. (2005). "Control of liver cell fate decision by a gradient of TGF beta signaling modulated by Onecut transcription factors." Genes Dev 19(16): 1849-54. Clotman, F., V. J. Lannoy, et al. (2002). "The onecut transcription factor HNF6 is required for normal development of the biliary tract." Development 129(8): 1819-28. Coffinier, C., L. Gresh, et al. (2002). "Bile system morphogenesis defects and liver dysfunction upon targeted deletion of HNF1beta." Development 129(8): 1829-38. Costa, R. H., D. R. Grayson, et al. (1989). "Multiple hepatocyte-enriched nuclear factors function in the regulation of transthyretin and alpha 1-antitrypsin genes." Mol Cell Biol 9(4): 1415-25. Coulouarn, C., C. Derambure, et al. (2005). "Global gene repression in hepatocellular carcinoma and fetal liver, and suppression of dudulin-2 mRNA as a possible marker for the cirrhosis-to138  tumor transition." J Hepatol 42(6): 860-9. Crawford, L. W., J. F. Foley, et al. (2010). "Histology atlas of the developing mouse hepatobiliary system with emphasis on embryonic days 9.5-18.5." Toxicol Pathol 38(6): 872-906. Crowe, A. J., L. Sang, et al. (1999). "Hepatocyte nuclear factor 3 relieves chromatin-mediated repression of the alpha-fetoprotein gene." J Biol Chem 274(35): 25113-20. Darlington, G. J., S. E. Ross, et al. (1998). "The role of C/EBP genes in adipocyte differentiation." J Biol Chem 273(46): 30057-60. Datson, N. A., J. van der Perk-de Jong, et al. (1999). "MicroSAGE: a modified procedure for serial analysis of gene expression in limited amounts of tissue." Nucleic Acids Res 27(5): 1300-7. Davies, D. (2007). Flow Cytometry: Principles and Applications (Chapter 11: Cell Sorting by Flow Cytometry), Humana Press Inc. de Boer, C. J., J. H. van Krieken, et al. (1999). "Expression of Ep-CAM in normal, regenerating, metaplastic, and neoplastic liver." J Pathol 188(2): 201-6. de Rave, S., H. W. Tilanus, et al. (2002). "The importance of orthotopic liver transplantation in acute hepatic failure." Transpl Int 15(1): 29-33. Decaens, T., C. Godard, et al. (2008). "Stabilization of beta-catenin affects mouse embryonic liver growth and hepatoblast fate." Hepatology 47(1): 247-58. DeLaForest, A., M. Nagaoka, et al. (2011). "HNF4A is essential for specification of hepatic progenitors from human pluripotent stem cells." Development 138(19): 4143-53. Dickson, P. W., G. J. Howlett, et al. (1985). "Rat transthyretin (prealbumin). Molecular cloning, nucleotide sequence, and gene expression in liver and brain." J Biol Chem 260(13): 8214-9. Dong, J., G. Feldmann, et al. (2007). "Elucidation of a universal size-control mechanism in Drosophila and mammals." Cell 130(6): 1120-33. Dorrell, C., L. Erker, et al. (2011). "Prospective isolation of a bipotential clonogenic liver progenitor cell in adult mice." Genes Dev 25(11): 1193-203. Dudas, J., M. Papoutsi, et al. (2004). "The homeobox transcription factor Prox1 is highly conserved in embryonic hepatoblasts and in adult and transformed hepatocytes, but is absent from bile duct epithelium." Anat Embryol (Berl) 208(5): 359-66. Duggan, D. J., M. Bittner, et al. (1999). "Expression profiling using cDNA microarrays." Nat Genet 21(1 Suppl): 10-4. Duncan, S. A., K. Manova, et al. (1994). "Expression of transcription factor HNF-4 in the extraembryonic endoderm, gut, and nephrogenic tissue of the developing mouse embryo: HNF-4 is a marker for primary endoderm in the implanting blastocyst." Proc Natl Acad Sci 139  U S A 91(16): 7598-602. Dyer, M. A., F. J. Livesey, et al. (2003). "Prox1 function controls progenitor cell proliferation and horizontal cell genesis in the mammalian retina." Nat Genet 34(1): 53-8. Eferl, R., M. Sibilia, et al. (1999). "Functions of c-Jun in liver and heart development." J Cell Biol 145(5): 1049-61. Fabian, M. R., N. Sonenberg, et al. (2010). "Regulation of mRNA translation and stability by microRNAs." Annu Rev Biochem 79: 351-79. Fasold, M., D. Langenberger, et al. (2011). "DARIO: a ncRNA detection and analysis tool for nextgeneration sequencing experiments." Nucleic Acids Res 39(Web Server issue): W112-7. Fassler, R. and M. Meyer (1995). "Consequences of lack of beta 1 integrin gene expression in mice." Genes Dev 9(15): 1896-908. Fausto, N. (2000). "Liver regeneration." J Hepatol 32(1 Suppl): 19-31. Ferguson, J. E., 3rd, Y. Wu, et al. (2007). "ASB4 is a hydroxylation substrate of FIH and promotes vascular differentiation via an oxygen-dependent mechanism." Mol Cell Biol 27(18): 640719. Festuccia, N. and I. Chambers (2011). "Quantification of pluripotency transcription factor levels in embryonic stem cells by flow cytometry." Curr Protoc Stem Cell Biol Chapter 1: Unit 1B 9. Floridon, C., C. H. Jensen, et al. (2000). "Does fetal antigen 1 (FA1) identify cells with regenerative, endocrine and neuroendocrine potentials? A study of FA1 in embryonic, fetal, and placental tissue and in maternal circulation." Differentiation 66(1): 49-59. Freeman, A. E., E. Engvall, et al. (1981). "Differentiation of fetal liver cells in vitro." Proc Natl Acad Sci U S A 78(6): 3659-63. Friedman, A. D., W. H. Landschulz, et al. (1989). "CCAAT/enhancer binding protein activates the promoter of the serum albumin gene in cultured hepatoma cells." Genes Dev 3(9): 1314-22. Friedman, J. R. and K. H. Kaestner (2006). "The Foxa family of transcription factors in development and metabolism." Cell Mol Life Sci 63(19-20): 2317-28. Furuyama, K., Y. Kawaguchi, et al. (2011). "Continuous cell supply from a Sox9-expressing progenitor zone in adult liver, exocrine pancreas and intestine." Nat Genet 43(1): 34-41. Garbett, K., P. J. Ebert, et al. (2008). "Immune transcriptome alterations in the temporal cortex of subjects with autism." Neurobiol Dis 30(3): 303-11. Garrison, W. D., M. A. Battle, et al. (2006). "Hepatocyte nuclear factor 4alpha is essential for embryonic development of the mouse colon." Gastroenterology 130(4): 1207-20.  140  Gaudreau, M. C., F. Heyd, et al. (2012). "Alternative splicing controlled by heterogeneous nuclear ribonucleoprotein L regulates development, proliferation, and migration of thymic pre-T cells." J Immunol 188(11): 5377-88. Ge, X., Q. Wu, et al. (2006). "A large quantity of novel human antisense transcripts detected by LongSAGE." Bioinformatics 22(20): 2475-9. Geerts, A. (2001). "History, heterogeneity, developmental biology, and functions of quiescent hepatic stellate cells." Semin Liver Dis 21(3): 311-35. Gerhard, D. S., L. Wagner, et al. (2004). "The status, quality, and expansion of the NIH full-length cDNA project: the Mammalian Gene Collection (MGC)." Genome Res 14(10B): 2121-7. Germain, L., M. J. Blouin, et al. (1988). "Biliary epithelial and hepatocytic cell lineage relationships in embryonic rat liver as determined by the differential expression of cytokeratins, alphafetoprotein, albumin, and cell surface-exposed components." Cancer Res 48(17): 4909-18. Ghosh, S. S., M. Takahashi, et al. (2000). "Liver-directed gene therapy: promises, problems and prospects at the turn of the century." J Hepatol 32(1 Suppl): 238-52. Giardine, B., C. Riemer, et al. (2005). "Galaxy: a platform for interactive large-scale genome analysis." Genome Res 15(10): 1451-5. Gillis, J. and P. Pavlidis (2011). "The impact of multifunctional genes on "guilt by association" analysis." PLoS One 6(2): e17258. Gonzalez, F. J. (2008). "Regulation of hepatocyte nuclear factor 4 alpha-mediated transcription." Drug Metab Pharmacokinet 23(1): 2-7. Gotea, V. and I. Ovcharenko (2008). "DiRE: identifying distant regulatory elements of co-expressed genes." Nucleic Acids Res 36(Web Server issue): W133-9. Greengard, O., M. Federman, et al. (1972). "Cytomorphometry of developing rat liver and its application to enzymic differentiation." J Cell Biol 52(2): 261-72. Gregoire, F. M., C. M. Smas, et al. (1998). "Understanding adipocyte differentiation." Physiol Rev 78(3): 783-809. Gresh, L., B. Bourachot, et al. (2005). "The SWI/SNF chromatin-remodeling complex subunit SNF5 is essential for hepatocyte differentiation." Embo J 24(18): 3313-24. Grozdanov, P. N., M. I. Yovchev, et al. (2006). "The oncofetal protein glypican-3 is a novel marker of hepatic progenitor/oval cells." Lab Invest 86(12): 1272-84. Gualdi, R., P. Bossard, et al. (1996). "Hepatic specification of the gut endoderm in vitro: cell signaling and transcriptional control." Genes Dev 10(13): 1670-82. Haque, M., C. H. Steinbrecher, et al. (2010). "Liver transplantation: Current status in British 141  Columbia." BC Medical Journal 52(4): 203-210. Haraguchi, Y., M. Takiguchi, et al. (1987). "Molecular cloning and nucleotide sequence of cDNA for human liver arginase." Proc Natl Acad Sci U S A 84(2): 412-5. Harbers, M. and P. Carninci (2005). "Tag-based approaches for transcriptome research and genome annotation." Nat Methods 2(7): 495-502. Hassan, A. S., J. Hou, et al. (2010). "Expression of two novel transcripts in the mouse definitive endoderm." Gene Expr Patterns 10(2-3): 127-34. Hauner, H. (2005). "Secretory factors from human adipose tissue and their functional role." Proc Nutr Soc 64(2): 163-9. Hawkes, C. and S. Kar (2004). "The insulin-like growth factor-II/mannose-6-phosphate receptor: structure, distribution and function in the central nervous system." Brain Res Brain Res Rev 44(2-3): 117-40. Hayhurst, G. P., Y. H. Lee, et al. (2001). "Hepatocyte nuclear factor 4alpha (nuclear receptor 2A1) is essential for maintenance of hepatic gene expression and lipid homeostasis." Mol Cell Biol 21(4): 1393-403. Heidelbaugh, J. J. and M. Bruderly (2006). "Cirrhosis and chronic liver failure: part I. Diagnosis and evaluation." Am Fam Physician 74(5): 756-62. Hentsch, B., I. Lyons, et al. (1996). "Hlx homeo box gene is essential for an inductive tissue interaction that drives expansion of embryonic liver and gut." Genes Dev 10(1): 70-9. Hilberg, F., A. Aguzzi, et al. (1993). "c-jun is essential for normal mouse development and hepatogenesis." Nature 365(6442): 179-81. Hirose, Y., T. Itoh, et al. (2009). "Hedgehog signal activation coordinates proliferation and differentiation of fetal liver progenitor cells." Exp Cell Res 315(15): 2648-57. Ho Sui, S. J., J. R. Mortimer, et al. (2005). "oPOSSUM: identification of over-represented transcription factor binding sites in co-expressed genes." Nucleic Acids Res 33(10): 3154-64. Hoffman, B. G., G. Robertson, et al. (2010). "Locus co-occupancy, nucleosome positioning, and H3K4me1 regulate the functionality of FOXA2-, HNF4A-, and PDX1-bound loci in islets and liver." Genome Res 20(8): 1037-51. Hoffman, B. G., B. Zavaglia, et al. (2008). "Identification of transcripts with enriched expression in the developing and adult pancreas." Genome Biol 9(6): R99. Hou, J., A. M. Charters, et al. (2007). "A systematic screen for genes expressed in definitive endoderm by Serial Analysis of Gene Expression (SAGE)." BMC Dev Biol 7: 92. Hu, M. and K. Polyak (2006). "Serial analysis of gene expression." Nat Protoc 1(4): 1743-60. 142  Huang da, W., B. T. Sherman, et al. (2007). "The DAVID Gene Functional Classification Tool: a novel biological module-centric algorithm to functionally analyze large gene lists." Genome Biol 8(9): R183. Huang, P., Z. He, et al. (2011). "Induction of functional hepatocyte-like cells from mouse fibroblasts by defined factors." Nature 475(7356): 386-9. Hunter, M. P., C. M. Wilson, et al. (2007). "The homeobox gene Hhex is essential for proper hepatoblast differentiation and bile duct morphogenesis." Dev Biol 308(2): 355-67. Ijzer, J. (2008). “Liver fibrosis and regeneration in dogs and cats: an immunohistochemical approach”, University of Utrecht. PhD: 159. Imren, S., M. E. Fabry, et al. (2004). "High-level beta-globin expression and preferred intragenic integration after lentiviral transduction of human cord blood stem cells." J Clin Invest 114(7): 953-62. Inoue, Y., G. P. Hayhurst, et al. (2002). "Defective ureagenesis in mice carrying a liver-specific disruption of hepatocyte nuclear factor 4alpha (HNF4alpha). HNF4alpha regulates ornithine transcarbamylase in vivo." J Biol Chem 277(28): 25257-65. Inoue, Y., A. M. Yu, et al. (2004). "Hepatocyte nuclear factor 4alpha is a central regulator of bile acid conjugation." J Biol Chem 279(4): 2480-9. Isern, J., S. T. Fraser, et al. (2008). "The fetal liver is a niche for maturation of primitive erythroid cells." Proc Natl Acad Sci U S A 105(18): 6662-7. Iwai, M., Y. Harada, et al. (2000). "Development of gap junctional channels and intercellular communication in rat liver during ontogenesis." J Hepatol 32(1): 11-8. Jochheim, A., A. Cieslak, et al. (2003). "Multi-stage analysis of differential gene expression in BALB/C mouse liver development by high-density microarrays." Differentiation 71(1): 6272. Jochheim, A., T. Hillemann, et al. (2004). "Quantitative gene expression profiling reveals a fetal hepatic phenotype of murine ES-derived hepatocytes." Int J Dev Biol 48(1): 23-9. Jochheim-Richter, A., U. Rudrich, et al. (2006). "Gene expression analysis identifies novel genes participating in early murine liver development and adult liver regeneration." Differentiation 74(4): 167-73. Jung, J., M. Zheng, et al. (1999). "Initiation of mammalian liver development from endoderm by fibroblast growth factors." Science 284(5422): 1998-2003. Kaestner, K. H., W. Knochel, et al. (2000). "Unified nomenclature for the winged helix/forkhead transcription factors." Genes Dev 14(2): 142-6. Kamiya, A., S. Kakinuma, et al. (2008). "Prospero-related homeobox 1 and liver receptor homolog 1 143  coordinately regulate long-term proliferation of murine fetal hepatoblasts." Hepatology 48(1): 252-64. Kamiya, A., T. Kinoshita, et al. (1999). "Fetal liver development requires a paracrine action of oncostatin M through the gp130 signal transducer." Embo J 18(8): 2127-36. Kanamura, S., K. Kanai, et al. (1985). "Quantitative analysis of development of mitochondrial ultrastructure in differentiating mouse hepatocytes during postnatal period." J Ultrastruct Res 93(3): 195-204. Kasahara, T., T. Miyazaki, et al. (2006). "Evaluation of methods for duration of preservation of RNA quality in rat liver used for transcriptome analysis." J Toxicol Sci 31(5): 509-19. Kellendonk, C., C. Opherk, et al. (2000). "Hepatocyte-specific expression of Cre recombinase." Genesis 26(2): 151-3. Khurana, S., A. K. Jaiswal, et al. (2010). "Hepatocyte nuclear factor-4alpha induces transdifferentiation of hematopoietic cells into hepatocytes." J Biol Chem 285(7): 4725-31. Khurana, S. and A. Mukhopadhyay (2008). "Hematopoietic progenitors from early murine fetal liver possess hepatic differentiation potential." Am J Pathol 173(6): 1818-27. Knowles, B. B., C. C. Howe, et al. (1980). "Human hepatocellular carcinoma cell lines secrete the major plasma proteins and hepatitis B surface antigen." Science 209(4455): 497-9. Kolterud, A., E. Wandzioch, et al. (2004). "Lhx2 is expressed in the septum transversum mesenchyme that becomes an integral part of the liver and the formation of these cells is independent of functional Lhx2." Gene Expr Patterns 4(5): 521-8. Krupczak-Hollis, K., X. Wang, et al. (2004). "The mouse Forkhead Box m1 transcription factor is essential for hepatoblast mitosis and development of intrahepatic bile ducts and vessels during liver morphogenesis." Dev Biol 276(1): 74-88. Kuimelis, R. G., K. J. Livak, et al. (1997). "Structural analogues of TaqMan probes for real-time quantitative PCR." Nucleic Acids Symp Ser(37): 255-6. Kurata, H., G. C. Mancini, et al. (1998). "Stem cell factor induces proliferation and differentiation of fetal progenitor cells in the mouse." Br J Haematol 101(4): 676-87. Kwon, A. T., D. J. Arenillas, et al. (2012). "oPOSSUM-3: Advanced Analysis of Regulatory Motif Over-Representation Across Genes or ChIP-Seq Datasets." G3 (Bethesda) 2(9): 987-1002. Kyrmizi, I., P. Hatzis, et al. (2006). "Plasticity and expanding complexity of the hepatic transcription factor network during liver development." Genes Dev 20(16): 2293-305. Lafarge, S., H. Hamzeh-Cognasse, et al. (2007). "A flow cytometry technique to study intracellular signals NF-kappaB and STAT3 in peripheral blood mononuclear cells." BMC Mol Biol 8: 64. 144  Lai, E. C. (2004). "Notch signaling: control of cell communication and cell fate." Development 131(5): 965-73. Lamb, P. and S. L. McKnight (1991). "Diversity and specificity in transcriptional regulation: the benefits of heterotypic dimerization." Trends Biochem Sci 16(11): 417-22. Lan, X., C. Adams, et al. (2011). "High resolution detection and analysis of CpG dinucleotides methylation using MBD-Seq technology." PLoS One 6(7): e22226. Lannoy, V. J., J. F. Decaux, et al. (2002). "Liver glucokinase gene expression is controlled by the onecut transcription factor hepatocyte nuclear factor-6." Diabetologia 45(8): 1136-41. Lazarevich, N. L. (2000). "Molecular mechanisms of alpha-fetoprotein gene expression." Biochemistry (Mosc) 65(1): 117-33. Lee, A. H., E. F. Scapa, et al. (2008). "Regulation of hepatic lipogenesis by the transcription factor XBP1." Science 320(5882): 1492-6. Lee, C. S., J. R. Friedman, et al. (2005). "The initiation of liver development is dependent on Foxa transcription factors." Nature 435(7044): 944-7. Lee, J. S., I. S. Chu, et al. (2004). "Classification and prediction of survival in hepatocellular carcinoma by gene expression profiling." Hepatology 40(3): 667-76. Lee, K., J. A. Villena, et al. (2003). "Inhibition of adipogenesis and development of glucose intolerance by soluble preadipocyte factor-1 (Pref-1)." J Clin Invest 111(4): 453-61. Lee, S., J. Bao, et al. (2005). "Detecting novel low-abundant transcripts in Drosophila." Rna 11(6): 939-46. Lemaigre, F. P. (2003). "Development of the biliary tract." Mech Dev 120(1): 81-7. Lemire, J. M., N. Shiojiri, et al. (1991). "Oval cell proliferation and the origin of small hepatocytes in liver injury induced by D-galactosamine." Am J Pathol 139(3): 535-52. Levesque, J. P., J. Hendy, et al. (2003). "Disruption of the CXCR4/CXCL12 chemotactic interaction during hematopoietic stem cell mobilization induced by GCSF or cyclophosphamide." J Clin Invest 111(2): 187-96. Li, A. C., C. J. Binder, et al. (2004). "Differential inhibition of macrophage foam-cell formation and atherosclerosis in mice by PPARalpha, beta/delta, and gamma." J Clin Invest 114(11): 156476. Li, B., V. Ruotti, et al. (2010). "RNA-Seq gene expression estimation with read mapping uncertainty." Bioinformatics 26(4): 493-500. Li, J., G. Ning, et al. (2000). "Mammalian hepatocyte differentiation requires the transcription factor HNF-4alpha." Genes Dev 14(4): 464-74. 145  Li, O., D. Vasudevan, et al. (2006). "High-level expression of DNA architectural factor HMGA2 and its association with nucleosomes in human embryonic stem cells." Genesis 44(11): 523-9. Li, T., J. Huang, et al. (2009). "Multi-stage analysis of gene expression and transcription regulation in C57/B6 mouse liver development." Genomics 93(3): 235-42. Lints, T. J., L. Hartley, et al. (1996). "Mesoderm-specific expression of the divergent homeobox gene Hlx during murine embryogenesis." Dev Dyn 205(4): 457-70. Liu, D., J. Fan, et al. (2010). "Quantitative analysis of miRNA expression in several developmental stages of human livers." Hepatol Res 40(8): 813-22. Liu, X. and J. E. Mertz (1995). "HnRNP L binds a cis-acting RNA sequence element that enables intron-dependent gene expression." Genes Dev 9(14): 1766-80. Lu, B., J. Yu, et al. (2009). "A novel approach to detect differentially expressed genes from countbased digital databases by normalizing with housekeeping genes." Genomics 94(3): 211-6. Ludtke, T. H., V. M. Christoffels, et al. (2009). "Tbx3 promotes liver bud expansion during mouse development by suppression of cholangiocyte differentiation." Hepatology 49(3): 969-78. Ma, Q., D. Jones, et al. (1999). "The chemokine receptor CXCR4 is required for the retention of B lineage and granulocytic precursors within the bone marrow microenvironment." Immunity 10(4): 463-71. Maass, T., I. Sfakianakis, et al. (2010). "Microarray-based gene expression analysis of hepatocellular carcinoma." Curr Genomics 11(4): 261-8. Mack, I., R. S. BelAiba, et al. (2009). "Functional analyses reveal the greater potency of preadipocytes compared with adipocytes as endothelial cell activator under normoxia, hypoxia, and TNFalpha exposure." Am J Physiol Endocrinol Metab 297(3): E735-48. Malig, R., C. Varela, et al. (2006). "Accurate and unambiguous tag-to-gene mapping in serial analysis of gene expression." BMC Bioinformatics 7: 487. Manganelli, R., S. Tyagi, et al. (2001). "Real Time PCR Using Molecular Beacons: A New Tool to Identify Point Mutations and to Analyze Gene Expression in Mycobacterium tuberculosis." Methods Mol Med 54: 295-310. Mansuroglu, T., J. Dudas, et al. (2009). "Hepatoblast and mesenchymal cell-specific geneexpression in fetal rat liver and in cultured fetal rat liver cells." Histochem Cell Biol 132(1): 11-9. Margagliotti, S., F. Clotman, et al. (2007). "The Onecut transcription factors HNF-6/OC-1 and OC-2 regulate early liver expansion by controlling hepatoblast migration." Dev Biol 311(2): 57989. Marrero, J. A., M. Kudo, et al. (2010). "The challenge of prognosis and staging for hepatocellular 146  carcinoma." Oncologist 15 Suppl 4: 23-33. Martinez Barbera, J. P., M. Clements, et al. (2000). "The homeobox gene Hex is required in definitive endodermal tissues for normal forebrain, liver and thyroid formation." Development 127(11): 2433-45. Matsumoto, K., H. Yoshitomi, et al. (2001). "Liver organogenesis promoted by endothelial cells prior to vascular function." Science 294(5542): 559-63. Matsumura, H., S. Reich, et al. (2003). "Gene expression analysis of plant host-pathogen interactions by SuperSAGE." Proc Natl Acad Sci U S A 100(26): 15718-23. McCarthy, K. F., M. L. Hale, et al. (1985). "Rat colony-forming unit spleen is OX7 positive, W3/13 positive, OX1 positive, and OX22 negative." Exp Hematol 13(9): 847-54. McKnight, K. D., J. Hou, et al. (2007). "Dynamic expression of thyrotropin-releasing hormone in the mouse definitive endoderm." Dev Dyn 236(10): 2909-17. Meehan, R. R., D. P. Barlow, et al. (1984). "Pattern of serum protein gene expression in mouse visceral yolk sac and foetal liver." Embo J 3(8): 1881-5. Millien, G., J. Beane, et al. (2008). "Characterization of the mid-foregut transcriptome identifies genes regulated during lung bud induction." Gene Expr Patterns 8(2): 124-39. Minguet, S., I. Cortegano, et al. (2003). "A population of c-Kit(low)(CD45/TER119)- hepatic cell progenitors of 11-day postcoitus mouse embryo liver reconstitutes cell-depleted liver organoids." J Clin Invest 112(8): 1152-63. Miyaoka, Y., M. Tanaka, et al. (2010). "A novel regulatory mechanism for Fgf18 signaling involving cysteine-rich FGF receptor (Cfr) and delta-like protein (Dlk)." Development 137(1): 159-67. Miyazaki, K., R. Takaki, et al. (1981). "Isolation and primary culture of adult human hepatocytes. Ultrastructural and functional studies." Cell Tissue Res 218(1): 13-21. Monaghan, A. P., K. H. Kaestner, et al. (1993). "Postimplantation expression patterns indicate a role for the mouse forkhead/HNF-3 alpha, beta and gamma genes in determination of the definitive endoderm, chordamesoderm and neuroectoderm." Development 119(3): 567-78. Moon, Y. S., C. M. Smas, et al. (2002). "Mice lacking paternally expressed Pref-1/Dlk1 display growth retardation and accelerated adiposity." Mol Cell Biol 22(15): 5585-92. Morgan, A. A., R. Chen, et al. (2012). "Clinical utility of sequence-based genotype compared with that derivable from genotyping arrays." J Am Med Inform Assoc 19(1e): e21-7. Morrissy, A. S. (2010). Bioinformatic analysis of cis-encoded antisense transcription. Medical Genetics. Vancouver, University of British Columbia. PhD: 158.  147  Morrissy, A. S., R. D. Morin, et al. (2009). "Next-generation tag sequencing for cancer gene expression profiling." Genome Res 19(10): 1825-35. Murakami, T., Y. Yasuda, et al. (1987). "Prealbumin gene expression during mouse development studied by in situ hybridization." Cell Differ 22(1): 1-9. Nakayama, M., K. Matsumoto, et al. (2006). "Id3 is important for proliferation and differentiation of the hepatoblasts during the chick liver development." Mech Dev 123(7): 580-90. Nesbitt, M. J., D. G. Moerman, et al. (2010). "Identifying novel genes in C. elegans using SAGE tags." BMC Mol Biol 11: 96. Nierhoff, D., A. Ogawa, et al. (2005). "Purification and characterization of mouse fetal liver epithelial cells with high in vivo repopulation capacity." Hepatology 42(1): 130-9. Nishino, J., I. Kim, et al. (2008). "Hmga2 promotes neural stem cell self-renewal in young but not old mice by reducing p16Ink4a and p19Arf Expression." Cell 135(2): 227-39. Nitou, M., K. Ishikawa, et al. (2000). "Immunohistochemical analysis of development of desminpositive hepatic stellate cells in mouse liver." J Anat 197 Pt 4: 635-46. O'Grady, J. G., S. W. Schalm, et al. (1993). "Acute liver failure: redefining the syndromes." Lancet 342(8866): 273-5. Odom, D. T., R. D. Dowell, et al. (2006). "Core transcriptional regulatory circuitry in human hepatocytes." Mol Syst Biol 2: 2006 0017. Odom, D. T., N. Zizlsperger, et al. (2004). "Control of pancreas and liver gene expression by HNF transcription factors." Science 303(5662): 1378-81. Oertel, M., A. Menthena, et al. (2008). "Purification of fetal liver stem/progenitor cells containing all the repopulation potential for normal adult rat liver." Gastroenterology 134(3): 823-32. Oikawa, T., A. Kamiya, et al. (2009). "Sall4 regulates cell fate decision in fetal hepatic stem/progenitor cells." Gastroenterology 136(3): 1000-11. Okabe, M., Y. Tsukahara, et al. (2009). "Potential hepatic stem cells reside in EpCAM+ cells of normal and injured mouse liver." Development 136(11): 1951-60. Onichtchouk, D., Y. G. Chen, et al. (1999). "Silencing of TGF-beta signalling by the pseudoreceptor BAMBI." Nature 401(6752): 480-5. Otu, H. H., K. Naxerova, et al. (2007). "Restoration of liver mass after injury requires proliferative and not embryonic transcriptional patterns." J Biol Chem 282(15): 11197-204. Palmada, M., S. Kanwal, et al. (2002). "c-jun is essential for sympathetic neuronal death induced by NGF withdrawal but not by p75 activation." J Cell Biol 158(3): 453-61.  148  Panadero, M., E. Herrera, et al. (2000). "Peroxisome proliferator-activated receptor-alpha expression in rat liver during postnatal development." Biochimie 82(8): 723-6. Papoutsi, M., J. Dudas, et al. (2007). "Gene regulation by homeobox transcription factor Prox1 in murine hepatoblasts." Cell Tissue Res 330(2): 209-20. Parviz, F., C. Matullo, et al. (2003). "Hepatocyte nuclear factor 4alpha controls the development of a hepatic epithelium and liver morphogenesis." Nat Genet 34(3): 292-6. Paulin, D. and Z. Li (2004). "Desmin: a major intermediate filament protein essential for the structural integrity and function of muscle." Exp Cell Res 301(1): 1-7. Pauws, E., A. H. van Kampen, et al. (2001). "Heterogeneity in polyadenylation cleavage sites in mammalian mRNA sequences: implications for SAGE analysis." Nucleic Acids Res 29(8): 1690-4. Pleasance, E. D., M. A. Marra, et al. (2003). "Assessment of SAGE in transcript identification." Genome Res 13(6A): 1203-15. Plumb-Rudewiez, N., F. Clotman, et al. (2004). "Transcription factor HNF-6/OC-1 inhibits the stimulation of the HNF-3alpha/Foxa1 gene by TGF-beta in mouse liver." Hepatology 40(6): 1266-74. Pons, F., M. Varela, et al. (2005). "Staging systems in hepatocellular carcinoma." HPB (Oxford) 7(1): 35-41. Pruitt, K. D., T. Tatusova, et al. (2007). "NCBI reference sequences (RefSeq): a curated nonredundant sequence database of genomes, transcripts and proteins." Nucleic Acids Res 35(Database issue): D61-5. Qu, X., E. Lam, et al. (2007). "Cited2, a coactivator of HNF4alpha, is essential for liver development." Embo J 26(21): 4445-56. Rapicavoli, N. A., E. M. Poth, et al. (2010). "The long noncoding RNA RNCR2 directs mouse retinal cell specification." BMC Dev Biol 10: 49. Reimold, A. M., A. Etkin, et al. (2000). "An essential role in liver development for transcription factor XBP-1." Genes Dev 14(2): 152-7. Rich, I. N. and B. Kubanek (1979). "The ontogeny of erythropoiesis in the mouse detected by the erythroid colony-forming technique. I. Hepatic and maternal erythropoiesis." J Embryol Exp Morphol 50: 57-74. Robertson, N., M. Oveisi-Fordorei, et al. (2007). "DiscoverySpace: an interactive data analysis application." Genome Biol 8(1): R6. Robinson, M. D. and A. Oshlack (2010). "A scaling normalization method for differential expression analysis of RNA-seq data." Genome Biol 11(3): R25. 149  Robinson, S. J., D. J. Cram, et al. (2004). "Maximizing the efficacy of SAGE analysis identifies novel transcripts in Arabidopsis." Plant Physiol 136(2): 3223-33. Roelandt, P., P. Sancho-Bru, et al. (2010). "Differentiation of rat multipotent adult progenitor cells to functional hepatocyte-like cells by mimicking embryonic liver development." Nat Protoc 5(7): 1324-36. Rogler, C. E., L. Levoci, et al. (2009). "MicroRNA-23b cluster microRNAs regulate transforming growth factor-beta/bone morphogenetic protein signaling and liver stem cell differentiation by targeting Smads." Hepatology 50(2): 575-84. Rogler, L. E. (1997). "Selective bipotential differentiation of mouse embryonic hepatoblasts in vitro." Am J Pathol 150(2): 591-602. Rollier, A., C. M. DiPersio, et al. (1993). "Regulation of albumin gene expression in hepatoma cells of fetal phenotype: dominant inhibition of HNF1 function and role of ubiquitous transcription factors." Mol Biol Cell 4(1): 59-69. Romanuik, T. L., G. Wang, et al. (2009). "Identification of novel androgen-responsive genes by sequencing of LongSAGE libraries." BMC Genomics 10: 476. Rosenbluh, J., D. Nijhawan, et al. (2011). "RMRP is a non-coding RNA essential for early murine development." PLoS One 6(10): e26270. Rossi, J. M., N. R. Dunn, et al. (2001). "Distinct mesodermal signals, including BMPs from the septum transversum mesenchyme, are required in combination for hepatogenesis from the endoderm." Genes Dev 15(15): 1998-2009. Ruzanov, P. and D. L. Riddle (2010). "Deep SAGE analysis of the Caenorhabditis elegans transcriptome." Nucleic Acids Res 38(10): 3252-62. Ryu, K. Y., R. Maehr, et al. (2007). "The mouse polyubiquitin gene UbC is essential for fetal liver development, cell-cycle progression and stress tolerance." Embo J 26(11): 2693-706. Saeed, A. I., V. Sharov, et al. (2003). "TM4: a free, open-source system for microarray data management and analysis." Biotechniques 34(2): 374-8. Saha, S., A. B. Sparks, et al. (2002). "Using the transcriptome to annotate the genome." Nat Biotechnol 20(5): 508-12. Sakai, H., Y. Tagawa, et al. (2010). "Isolation and characterization of portal branch ligationstimulated Hmga2-positive bipotent hepatic progenitor cells." Biochem Biophys Res Commun 403(3-4): 298-304. Sakajiri, S., J. O'Kelly, et al. (2005). "Dlk1 in normal and abnormal hematopoiesis." Leukemia 19(8): 1404-10. Sandelin, A., W. Alkema, et al. (2004). "JASPAR: an open-access database for eukaryotic 150  transcription factor binding profiles." Nucleic Acids Res 32(Database issue): D91-4. Sano, K., Y. Miyamoto, et al. (2010). "Survival signals of hepatic stellate cells in liver regeneration are regulated by glycosylation changes in rat vitronectin, especially decreased sialylation." J Biol Chem 285(23): 17301-9. Santangelo, L., A. Marchetti, et al. (2011). "The stable repression of mesenchymal program is required for hepatocyte identity: a novel role for hepatocyte nuclear factor 4alpha." Hepatology 53(6): 2063-74. Sasaki, K. and G. Matsumura (1986). "Haemopoietic cells of yolk sac and liver in the mouse embryo: a light and electron microscopical study." J Anat 148: 87-97. Sasik, R., N. Iranfar, et al. (2002). "Extracting transcriptional events from temporal gene expression patterns during Dictyostelium development." Bioinformatics 18(1): 61-6. Schmid, P. and W. A. Schulz (1990). "Coexpression of the c-myc protooncogene with alphafetoprotein and albumin in fetal mouse liver." Differentiation 45(2): 96-102. Schmidt, C., F. Bladt, et al. (1995). "Scatter factor/hepatocyte growth factor is essential for liver development." Nature 373(6516): 699-702. Schrem, H., J. Klempnauer, et al. (2004). "Liver-enriched transcription factors in liver function and development. Part II: the C/EBPs and D site-binding protein in cell cycle control, carcinogenesis, circadian gene regulation, liver regeneration, apoptosis, and liver-specific gene regulation." Pharmacol Rev 56(2): 291-330. Sekine, S., R. Ogawa, et al. (2011). "Hepatomas with activating Ctnnb1 mutations in 'Ctnnb1deficient' livers: a tricky aspect of a conditional knockout mouse model." Carcinogenesis 32(4): 622-8. Sekine, S., R. Ogawa, et al. (2009). "Dicer is required for proper liver zonation." J Pathol 219(3): 365-72. Sekiya, S. and A. Suzuki (2011). "Direct conversion of mouse fibroblasts to hepatocyte-like cells by defined factors." Nature 475(7356): 390-3. Sellem, C. H., M. Frain, et al. (1984). "Differential expression of albumin and alpha-fetoprotein genes in fetal tissues of mouse and rat." Dev Biol 102(1): 51-60. Sen, G. L., J. A. Reuter, et al. (2010). "DNMT1 maintains progenitor function in self-renewing somatic tissue." Nature 463(7280): 563-7. Septer, S., G. Edwards, et al. (2012). "Yes-associated protein is involved in proliferation and differentiation during postnatal liver development." Am J Physiol Gastrointest Liver Physiol 302(5): G493-503. Shah, S. A., G. A. Levy, et al. (2006). "Adult-to-adult living donor liver transplantation." Can J 151  Gastroenterol 20(5): 339-43. Shell, S., S. M. Park, et al. (2007). "Let-7 expression defines two differentiation stages of cancer." Proc Natl Acad Sci U S A 104(27): 11400-5. Shin, S., G. Walton, et al. (2011). "Foxl1-Cre-marked adult hepatic progenitors have clonogenic and bilineage differentiation potential." Genes Dev 25(11): 1185-92. Shiojiri, N. (1981). "Enzymo- and immunocytochemical analyses of the differentiation of liver cells in the prenatal mouse." J Embryol Exp Morphol 62: 139-52. Shiojiri, N. (1984). "The origin of intrahepatic bile duct cells in the mouse." J Embryol Exp Morphol 79: 25-39. Shiojiri, N., J. M. Lemire, et al. (1991). "Cell lineages and oval cell progenitors in rat liver development." Cancer Res 51(10): 2611-20. Si-Tayeb, K., F. P. Lemaigre, et al. (2010). "Organogenesis and development of the liver." Dev Cell 18(2): 175-89. Siddiqui, A. S., J. Khattra, et al. (2005). "A mouse atlas of gene expression: large-scale digital geneexpression profiles from precisely defined developing C57BL/6J mouse tissues and cells." Proc Natl Acad Sci U S A 102(51): 18485-90. Sigal, S. H., S. Brill, et al. (1994). "Characterization and enrichment of fetal rat hepatoblasts by immunoadsorption ("panning") and fluorescence-activated cell sorting." Hepatology 19(4): 999-1006. Sladek, F. M., W. M. Zhong, et al. (1990). "Liver-enriched transcription factor HNF-4 is a novel member of the steroid hormone receptor superfamily." Genes Dev 4(12B): 2353-65. Smas, C. M. and H. S. Sul (1993). "Pref-1, a protein containing EGF-like repeats, inhibits adipocyte differentiation." Cell 73(4): 725-34. Sosa-Pineda, B., J. T. Wigle, et al. (2000). "Hepatocyte migration during liver development requires Prox1." Nat Genet 25(3): 254-5. Soto-Gutierrez, A., N. Navarro-Alvarez, et al. (2007). "Differentiation of mouse embryonic stem cells to hepatocyte-like cells by co-culture with human liver nonparenchymal cell lines." Nat Protoc 2(2): 347-56. Spear, B. T., L. Jin, et al. (2006). "Transcriptional control in the mammalian liver: liver development, perinatal repression, and zonal gene regulation." Cell Mol Life Sci 63(24): 2922-38. Spector, D. L. and R. D. Goldman (2006). Basic Methods in Microscopy, CSHL Press. Stenvers, K. L., M. L. Tursky, et al. (2003). "Heart and liver defects and reduced transforming 152  growth factor beta2 sensitivity in transforming growth factor beta type III receptor-deficient embryos." Mol Cell Biol 23(12): 4371-85. Stevenson, E. J., P. G. Giresi, et al. (2003). "Global analysis of gene expression patterns during disuse atrophy in rat skeletal muscle." J Physiol 551(Pt 1): 33-48. Strain, A. J. and J. M. Neuberger (2002). "A bioartificial liver--state of the art." Science 295(5557): 1005-9. Sul, H. S. (2009). "Minireview: Pref-1: role in adipogenesis and mesenchymal cell fate." Mol Endocrinol 23(11): 1717-25. Sun, M., G. Zhou, et al. (2004). "SAGE is far more sensitive than EST for detecting low-abundance transcripts." BMC Genomics 5(1): 1. Suzuki, A., S. Sekiya, et al. (2008). "Tbx3 controls the fate of hepatic progenitor cells in liver development by suppressing p19ARF expression." Development 135(9): 1589-95. Suzuki, A., Y. Zheng, et al. (2000). "Flow-cytometric separation and enrichment of hepatic progenitor cells in the developing mouse liver." Hepatology 32(6): 1230-9. Suzuki, A., Y. W. Zheng, et al. (2002). "Clonal identification and characterization of self-renewing pluripotent stem cells in the developing liver." J Cell Biol 156(1): 173-84. Szpirer, C. and J. Szpirer (1975). "A mouse hepatoma cell line which secretes several serum proteins including albumin and alpha-foetoprotein." Differentiation 4(2): 85-91. Takabe, Y., S. Yagi, et al. (2012). "Immunomagnetic exclusion of E-cadherin-positive hepatoblasts in fetal mouse liver cell cultures impairs morphogenesis and gene expression of sinusoidal endothelial cells." J Anat 221(3): 229-39. Tan, X., Y. Yuan, et al. (2008). "Beta-catenin deletion in hepatoblasts disrupts hepatic morphogenesis and survival during mouse development." Hepatology 47(5): 1667-79. Tanaka, M., M. Okabe, et al. (2009). "Mouse hepatoblasts at distinct developmental stages are characterized by expression of EpCAM and DLK1: drastic change of EpCAM expression during liver development." Mech Dev 126(8-9): 665-76. Tanimizu, N. and A. Miyajima (2004). "Notch signaling controls hepatoblast differentiation by altering the expression of liver-enriched transcription factors." J Cell Sci 117(Pt 15): 316574. Tanimizu, N., A. Miyajima, et al. (2007). "Liver progenitor cells develop cholangiocyte-type epithelial polarity in three-dimensional culture." Mol Biol Cell 18(4): 1472-9. Tanimizu, N., A. Miyajima, et al. (2009). "Liver progenitor cells fold up a cell monolayer into a double-layered structure during tubular morphogenesis." Mol Biol Cell 20(9): 2486-94.  153  Tanimizu, N., M. Nishikawa, et al. (2003). "Isolation of hepatoblasts based on the expression of Dlk/Pref-1." J Cell Sci 116(Pt 9): 1775-86. Tanimizu, N., H. Saito, et al. (2004). "Long-term culture of hepatic progenitors derived from mouse Dlk+ hepatoblasts." J Cell Sci 117(Pt 26): 6425-34. Tanimizu, N., T. Tsujimura, et al. (2004). "Expression of Dlk/Pref-1 defines a subpopulation in the oval cell compartment of rat liver." Gene Expr Patterns 5(2): 209-18. Taraviras, S., A. P. Monaghan, et al. (1994). "Characterization of the mouse HNF-4 gene and its expression during mouse embryogenesis." Mech Dev 48(2): 67-79. Theiler, K. (1989). The House Mouse: Atlas of Embryonic Development. New York, SpringerVerlag. Trayhurn, P. (2005). "The biology of obesity." Proc Nutr Soc 64(1): 31-8. Trowbridge, J. J., J. W. Snow, et al. (2009). "DNA methyltransferase 1 is essential for and uniquely regulates hematopoietic stem and progenitor cells." Cell Stem Cell 5(4): 442-9. Uchida, N., T. Fujisaki, et al. (2001). "Transplantable hematopoietic stem cells in human fetal liver have a CD34(+) side population (SP)phenotype." J Clin Invest 108(7): 1071-7. Vagima, Y., K. Lapid, et al. (2011). "Pathways implicated in stem cell migration: the SDF-1/CXCR4 axis." Methods Mol Biol 750: 277-89. Van Lint, P. and C. Libert (2007). "Chemokine and cytokine processing by matrix metalloproteinases and its effect on leukocyte migration and inflammation." J Leukoc Biol 82(6): 1375-81. van Ruissen, F., J. M. Ruijter, et al. (2005). "Evaluation of the similarity of gene expression data estimated with SAGE and Affymetrix GeneChips." BMC Genomics 6: 91. Vanhaecke, T. and V. Rogiers (2006). "Hepatocyte cultures in drug metabolism and toxicological research and testing." Methods Mol Biol 320: 209-27. Velculescu, V. E., L. Zhang, et al. (1995). "Serial analysis of gene expression." Science 270(5235): 484-7. von Bubnoff, D., H. Bausinger, et al. (2004). "Human epidermal langerhans cells express the immunoregulatory enzyme indoleamine 2,3-dioxygenase." J Invest Dermatol 123(2): 298304. Vosough, M., M. Moslem, et al. (2011). "Cell-based therapeutics for liver disorders." Br Med Bull 100: 157-72. Vrljicak, P. (2010). “Genomic analysis of embryonic heart development in the mouse.” Vancouver, University of British Columbia. PhD: 174. 154  Vrljicak, P., A. C. Chang, et al. (2010). "Genomic analysis distinguishes phases of early development of the mouse atrio-ventricular canal." Physiol Genomics 40(3): 150-7. Vroman, B. and N. F. LaRusso (1996). "Development and characterization of polarized primary cultures of rat intrahepatic bile duct epithelial cells." Lab Invest 74(1): 303-13. Waddell, J. N., P. Zhang, et al. (2010). "Dlk1 is necessary for proper skeletal muscle development and regeneration." PLoS One 5(11): e15055. Wahl, M. B., U. Heinzmann, et al. (2005). "LongSAGE analysis revealed the presence of a large number of novel antisense genes in the mouse genome." Bioinformatics 21(8): 1389-92. Wamstad, J. A., J. M. Alexander, et al. (2012). "Dynamic and Coordinated Epigenetic Regulation of Developmental Transitions in the Cardiac Lineage." Cell. Wang, Y., K. A. Kim, et al. (2006). "Pref-1, a preadipocyte secreted factor that inhibits adipogenesis." J Nutr 136(12): 2953-6. Watanabe, N., M. Tanaka, et al. (2007). "Tim2 is expressed in mouse fetal hepatocytes and regulates their differentiation." Hepatology 45(5): 1240-9. Watt, A. J., R. Zhao, et al. (2007). "Development of the mammalian liver and ventral pancreas is dependent on GATA4." BMC Dev Biol 7: 37. Wei, C. L., P. Ng, et al. (2004). "5' Long serial analysis of gene expression (LongSAGE) and 3' LongSAGE for transcriptome characterization and genome annotation." Proc Natl Acad Sci U S A 101(32): 11701-6. Weinstein, M., S. P. Monga, et al. (2001). "Smad proteins and hepatocyte growth factor control parallel regulatory pathways that converge on beta1-integrin to promote normal liver development." Mol Cell Biol 21(15): 5122-31. Wellen, K. E. and G. S. Hotamisligil (2005). "Inflammation, stress, and diabetes." J Clin Invest 115(5): 1111-9. Westmacott, A., Z. D. Burke, et al. (2006). "C/EBPalpha and C/EBPbeta are markers of early liver development." Int J Dev Biol 50(7): 653-7. Wilson, J. W., C. S. Groat, et al. (1963). "Histogenesis Of The Liver." Ann N Y Acad Sci 111: 8-24. Wingender, E., X. Chen, et al. (2000). "TRANSFAC: an integrated system for gene expression regulation." Nucleic Acids Res 28(1): 316-9. Winter, N., R. Nimzyk, et al. (2011). "Chromatin immunoprecipitation to analyze DNA binding sites of HMGA2." PLoS One 6(4): e18837. Wisse, E., D. Luo, et al. (1997). "On the function of pit cells, the liver-specific natural killer cells." Semin Liver Dis 17(4): 265-86. 155  Wojciechowicz, K., E. Markiewicz, et al. (2008). "C/EBPalpha identifies differentiating preadipocytes around hair follicles in foetal and neonatal rat and mouse skin." Exp Dermatol 17(8): 675-80. Wolber, F. M., E. Leonard, et al. (2002). "Roles of spleen and liver in development of the murine hematopoietic system." Exp Hematol 30(9): 1010-9. Wolfrum, C., E. Asilmaz, et al. (2004). "Foxa2 regulates lipid metabolism and ketogenesis in the liver during fasting and in diabetes." Nature 432(7020): 1027-32. Wong, M. L. and J. F. Medrano (2005). "Real-time PCR for mRNA quantitation." Biotechniques 39(1): 75-85. Wu, Q., M. Kawahara, et al. (2008). "Synergistic role of Igf2 and Dlk1 in fetal liver development and hematopoiesis in bi-maternal mice." J Reprod Dev 54(3): 177-82. Wu, Z. J., C. A. Meyer, et al. (2010). "Gene expression profiling of human breast tissue samples using SAGE-Seq." Genome Res 20(12): 1730-9. Xiang, X., K. F. Benson, et al. (1990). "Mini-mouse: disruption of the pygmy locus in a transgenic insertional mutant." Science 247(4945): 967-9. Xie, X., H. Liao, et al. (2009). "Down-regulation of hepatic HNF4alpha gene expression during hyperinsulinemia via SREBPs." Mol Endocrinol 23(4): 434-43. Xu, B. and Y. Huang (2009). "Histone H2a mRNA interacts with Lin28 and contains a Lin28dependent posttranscriptional regulatory element." Nucleic Acids Res 37(13): 4256-63. Yamasaki, H., A. Sada, et al. (2006). "Suppression of C/EBPalpha expression in periportal hepatoblasts may stimulate biliary cell differentiation through increased Hnf6 and Hnf1b expression." Development 133(21): 4233-43. Yanai, H., K. Nakamura, et al. (2010). "Dlk-1, a cell surface antigen on foetal hepatic stem/progenitor cells, is expressed in hepatocellular, colon, pancreas and breast carcinomas at a high frequency." J Biochem 148(1): 85-92. Yang, C. H., L. Y. Chuang, et al. (2010). "hSAGEing: an improved SAGE-based software for identification of human tissue-specific or common tumor markers and suppressors." PLoS One 5(12): e14369. Yang, F., T. Agulian, et al. (2004). "Developmental regulation of galactokinase in suckling mouse liver by the Egr-1 transcription factor." Pediatr Res 55(5): 822-9. Yang, W. J., D. D. Yang, et al. (2005). "Dicer is required for embryonic angiogenesis during mouse development." J Biol Chem 280(10): 9330-5. Yokoyama, S., M. Hashimoto, et al. (2008). "Dynamic gene expression of Lin-28 during embryonic development in mouse and chicken." Gene Expr Patterns 8(3): 155-60. 156  Yu, Y., J. Ping, et al. (2010). "A comparative analysis of liver transcriptome suggests divergent liver function among human, mouse and rat." Genomics 96(5): 281-9. Yuen, M. F. and C. L. Lai (2005). "Serological markers of liver cancer." Best Pract Res Clin Gastroenterol 19(1): 91-9. Zaret, K. (1998). "Early liver differentiation: genetic potentiation and multilevel growth control." Curr Opin Genet Dev 8(5): 526-31. Zaret, K. S. (2002). "Regulatory phases of early liver development: paradigms of organogenesis." Nat Rev Genet 3(7): 499-512. Zaret, K. S., J. Watts, et al. (2008). "Pioneer factors, genetic competence, and inductive signaling: programming liver and pancreas progenitors from the endoderm." Cold Spring Harb Symp Quant Biol 73: 119-26. Zhang, C. C. and H. F. Lodish (2004). "Insulin-like growth factor 2 expressed in a novel fetal liver cell population is a growth factor for hematopoietic stem cells." Blood 103(7): 2513-21. Zhao, R. and S. A. Duncan (2005). "Embryonic development of the liver." Hepatology 41(5): 95667. Zheng, K., X. Wu, et al. (2009). "The pluripotency factor LIN28 marks undifferentiated spermatogonia in mouse." BMC Dev Biol 9: 38. Zhou, Y., J. Lee, et al. (2011). "Regulation of glucose homeostasis through a XBP-1-FoxO1 interaction." Nat Med 17(3): 356-65. Zhu, N. L., K. Asahina, et al. (2012). "Hepatic stellate cell-derived delta-like homolog 1 (DLK1) protein in liver regeneration." J Biol Chem 287(13): 10355-67. Zorn, A. M. (2008). "Liver development." StemBook.  157  Appendix Appendix A Cluster A genes and their expression in the five Tag-seq libraries (in tags per million).  Symbol 2310022B05Rik 2410022L05Rik 2610029G23Rik 2810030E01Rik 2810408I11Rik 2900062L11Rik 4921524J17Rik 5830415L20Rik 6230409E13Rik 6330403K07Rik Agpat1 Apitd1 Arglu1 Asb4 Atp2a2 Atp6v0e Bambi Barx1 Bat2d BC003885 Bzw2 C80913 Car4 Cbfb Ccdc124 Ccdc34 Ccnd2 Cd24a  E10.5 DLK1+ 22 22 30 24 14 78 16 26 12 31 11 15 21 549 137 119 63 12 32 31 100 25 108 79 27 118 162 48  E12.5 DLK1+ 4 2 5 0 0 2 0 0 0 0 0 0 2 8 20 13 6 0 6 2 14 3 4 11 3 12 8 4  E14.5 DLK1+ 2 3 3 5 1 0 2 3 0 2 1 1 1 14 12 14 4 0 3 0 10 3 4 11 4 22 3 2  E16.5 DLK1+ 9 4 5 6 2 34 9 13 1 18 4 2 11 60 50 53 12 4 12 9 15 10 0 37 6 42 54 17  adult liver 0 0 0 0 0 0 0 2 0 0 0 0 0 0 8 5 5 0 2 0 0 0 0 4 0 0 0 0  Cd320 Cenpq Cep55 Clip3 Cnot7 Cnot8 Cops7b Cpxm1 Crispld2 Csrp2 Ctxn1 Cyfip2 D030056L22Rik Dck Ddah2 Dek Dennd2a Dhrs13 Dhx9 Dnajc12 Dnmt1 Dscc1 Dtl Eed Eef1b2 Eid2 Eif3m Emb Epcam Etaa1 Ezh2 Fam83d Fbxo5 Fkbp3 Flnc Flrt3  21 12 31 15 12 23 12 25 12 59 16 15 51 50 12 63 29 13 81 40 71 12 21 75 17 19 97 234 16 11 42 11 74 22 14 12  1 2 3 0 1 0 0 0 0 7 0 2 8 4 0 4 4 0 8 2 4 0 0 6 0 0 0 13 0 0 4 0 11 0 0 0  2 0 4 0 0 0 1 2 0 2 0 2 2 5 0 9 1 2 16 7 6 0 2 5 1 0 4 2 0 0 3 2 10 3 0 0  6 5 15 8 6 11 1 3 1 7 4 4 9 17 4 31 7 3 30 12 18 0 5 20 5 5 19 14 0 4 10 3 16 5 5 3  3 0 0 0 0 2 0 0 0 6 0 0 0 2 0 0 2 0 7 2 0 0 0 5 0 0 2 0 0 0 0 0 0 0 0 0  158  Frat2 Frem2 Fxc1 Fzd2 Gar1 Gata6 Gjc1 Glmn Gm1673 Hdac2 Hist3h2ba Hmga2 Hn1l Hnrnpl Hsd11b2 Igfbp5 Itgb3bp Jam3 Jarid2 Kitl Klc1 Lin28 Llph LOC100045887 LOC100047028 Lpar2 Lyrm4 Maml1 Map4k4 Marcksl1 Mcm8 Mdc1 Med17 Med19 Med30 Mns1  18 11 23 22 19 30 17 47 16 164 14 520 74 1268 20 22 30 28 12 98 24 93 14 35 29 27 11 10 90 391 14 37 14 25 39 12  2 0 0 0 0 2 2 2 0 19 2 39 7 217 0 0 5 1 1 13 1 17 2 0 0 3 2 2 15 18 2 4 0 2 2 0  3 0 3 0 0 3 0 3 0 14 1 22 4 165 0 0 3 0 1 5 2 10 2 2 2 0 1 1 14 5 1 5 3 2 4 1  2 2 10 5 2 11 6 16 3 74 1 18 18 546 5 7 5 5 5 39 10 1 3 7 4 0 6 5 32 78 4 12 4 8 7 4  0 0 0 0 0 0 0 0 0 4 0 0 0 71 0 0 0 0 0 0 0 0 0 2 0 0 0 0 2 12 0 0 0 2 2 0  Mrpl20 Mrpl43 Mrpl45 Nap1l1 Nmral1 Nob1 Nsmce2 Nup133 Nup153 Nup93 P4ha1 Papola Pebp1 Pftk1 Plekha3 Polr3g Ppil5 Ppp1r1a Ppp1r7 Prmt5 Prtg Psmc6 Psrc1 Ptpn13 Pus3 Rabggtb Rabl2a Rad51 Rai12 Ran Ranbp1 Rbbp4 Rcn2 Rcor2 Rilpl1 Robo1  127 42 18 323 34 23 36 30 23 31 21 49 167 22 11 15 16 16 19 262 39 72 13 16 13 110 10 53 31 19 261 146 78 21 14 209  21 6 0 34 0 0 6 5 0 4 1 4 5 4 0 0 0 0 0 36 5 5 1 4 0 13 0 8 4 3 22 26 3 4 2 39  13 5 1 17 7 3 6 5 2 3 1 1 5 0 0 3 3 0 4 43 0 8 2 0 2 19 1 9 3 0 6 11 6 2 0 14  20 14 10 61 8 5 14 11 8 6 11 19 37 9 6 5 3 0 6 51 1 31 6 7 3 42 3 12 12 4 30 14 30 1 4 104  12 5 0 20 0 0 0 0 0 0 0 0 6 0 0 0 0 0 0 0 0 4 0 0 0 8 0 0 0 0 0 4 4 0 0 0  159  Rprd1b Rps27l Rrp1 Rsl1d1 Sall2 Sdf2l1 Sepn1 Sfrs4 Sfrs7 Shisa3 Slc39a10 Smad2 Smtnl2 Snrpb2 Snrpd3 Sox11 Sox13 Sox9 Spata5 Sulf2 Sumo1 Tbpl1 Tcf7 Tead2 Tes Tgfb2 Thoc4 Tpbg Tpm1 Traf4 Traip Trove2 Tspan6 Ttyh3 Twist1 Uba2  15 438 21 178 13 67 19 14 311 14 17 25 14 252 84 138 11 20 10 43 30 25 21 69 19 13 43 14 38 17 10 13 51 11 12 60  0 43 0 24 0 11 1 2 50 0 0 4 1 39 11 6 0 0 0 0 0 2 2 3 0 0 3 0 5 0 0 2 9 2 0 2  2 49 2 15 0 5 3 0 21 0 0 5 0 37 13 0 1 0 0 0 0 2 0 4 3 0 3 0 2 3 1 1 3 0 0 4  6 114 8 28 2 11 7 5 64 1 5 11 7 72 23 8 3 4 2 4 13 11 1 6 3 1 8 2 18 7 1 3 6 4 1 15  0 14 2 15 0 5 0 0 12 0 0 0 0 20 5 0 0 0 0 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2  Ube2q2 Wdr85 Whsc2 Yap1 Ythdf1 Zfp106 Zfp280c  21 12 37 61 42 17 26  2 1 4 7 5 0 5  1 2 7 9 8 1 1  10 3 11 17 15 5 10  2 0 0 0 0 0 0  160  Appendix B Cluster B genes and their expression in the five Tag-seq libraries (in tags per million).  Symbol 0610007P08Rik 0610007P22Rik 0610009B22Rik 0610011F06Rik 0610011L14Rik 0610038F07Rik 1110021L09Rik 1110038F14Rik 1110057K04Rik 1110059G10Rik 1200003C05Rik 1200009F10Rik 1200011I18Rik 1200016B10Rik 1500011K16Rik 1500035H01Rik 1700123O20Rik 1810014B01Rik 1810035L17Rik 2010012O05Rik 2210012G02Rik 2210018M11Rik 2210411K11Rik 2310001A20Rik 2310007F21Rik 2310036O22Rik 2310039H08Rik  E10.5 DLK1+ 11 34 58 62 26 46 20 13 15 18 25 16 15 34 65 11 89 12 227 57 18 12 23 24 10 88 11  E12.5 DLK1+ 5 19 32 29 15 15 8 2 4 7 2 6 2 19 44 10 59 6 160 20 9 8 3 8 5 20 6  E14.5 DLK1+ 1 15 40 20 14 28 6 4 6 3 7 2 12 15 24 8 30 6 112 19 8 2 10 10 4 20 5  E16.5 DLK1+ 8 15 27 39 6 20 12 7 12 10 16 11 10 29 22 12 23 7 58 32 14 7 12 19 7 22 7  adult liver 4 17 47 32 22 14 4 3 9 8 9 3 10 21 12 7 57 3 63 43 8 3 4 9 3 19 6  2410004B18Rik 2410016O06Rik 2510003E04Rik 2610110G12Rik 2810405K02Rik 4632411B12Rik 4632428N05Rik 4732418C07Rik 4833439L19Rik 4930452B06Rik 4930455C21Rik 4932417H02Rik 4933407N01Rik 5330431N19Rik 5730437N04Rik 6330578E17Rik 6530401N04Rik 9530068E07Rik A130010J15Rik Aarsd1 Abca1 Abcb7 Abce1 Abcf1 Abhd10 Abhd12 Abhd3 Acads Acbd6 Acn9 Actr1a Actr2  45 30 14 15 14 14 7 18 36 12 52 14 10 12 19 41 18 126 12 22 139 23 176 19 38 17 38 38 34 19 22 138  25 19 2 7 6 2 0 10 9 8 9 4 2 0 8 30 3 48 7 19 63 7 135 6 25 7 23 18 16 9 15 82  16 20 3 4 6 3 2 9 4 4 15 7 4 3 5 30 4 39 3 5 88 6 80 4 14 6 34 24 18 12 15 60  25 4 7 9 12 16 10 18 17 5 38 5 8 4 8 21 7 60 4 3 166 22 55 9 16 10 53 19 14 7 11 90  8 19 9 12 7 5 4 12 6 4 12 9 4 2 7 26 3 19 4 11 97 15 72 3 10 13 28 18 18 15 7 79  161  Actr6 Actr8 Adam10 Adi1 Adipor1 Adnp Ado Ahcyl1 Ahcyl2 AI314976 Aifm1 Akap8 Alg1 Alg14 Alkbh7 Als2 Ankib1 Anp32a Aof1 Ap2s1 Ap4b1 Aqr Arcn1 Arf4 Arid1b Arih1 Arl13b Arl6ip1 Arl6ip5 Arpc1a Atad1 Atad3a  16 24 54 15 56 32 13 60 9 71 44 78 11 18 38 28 16 35 10 44 28 27 124 439 30 48 13 499 35 66 154 28  6 15 34 2 11 15 2 27 3 40 23 52 7 9 20 20 0 23 9 39 12 14 56 124 10 31 12 210 13 21 85 15  7 19 16 5 18 8 6 41 2 27 10 44 3 3 14 22 6 17 5 18 26 6 110 176 8 14 7 237 9 11 57 15  13 8 36 17 39 16 5 44 5 14 28 49 5 8 11 22 11 17 13 23 17 10 73 376 23 26 19 323 18 29 103 11  6 10 9 11 18 7 6 41 11 23 14 30 5 4 23 17 8 13 7 19 17 9 63 167 17 15 16 126 34 36 71 14  Atf4 Atf6b Atg12 Atg3 Atp13a1 Atp1b3 Atp5g3 Atp6ap1 Atp6v0b Atp6v1d Atp6v1h Atp9b Atrx Aurkaip1 Axin2 B4galt3 B4galt5 B630005N14Rik Bag4 Bap1 Bat2l BC003940 BC005624 BC088983 Bcap31 Bcl7b Bcor Bet1 Blzf1 Bmpr1a Bms1 Bnip3l  161 20 16 15 11 130 99 47 20 30 42 11 42 37 13 15 78 56 24 32 21 42 21 14 73 12 13 46 16 76 57 68  61 8 12 6 9 33 36 21 6 16 32 8 8 7 2 4 44 23 6 0 4 11 5 4 23 5 12 29 4 30 31 16  51 8 5 3 5 37 43 12 8 23 22 6 10 5 3 3 25 18 9 6 8 14 8 1 21 8 8 22 10 22 15 27  109 4 14 13 5 63 57 25 17 29 27 9 44 22 0 6 37 34 6 10 13 29 10 8 42 8 11 46 10 45 19 43  108 5 10 10 7 20 47 24 9 17 33 6 25 14 3 5 15 15 8 7 7 8 5 2 27 7 8 38 5 18 12 37  162  Bptf Brd9 Bri3bp Bub3 C030044B11Rik C1d C330023M02Rik C87436 Calm1 Camk2n1 Camsap1 Casp6 Ccdc104 Ccdc47 Ccdc97 Ccnh Cct5 Cd164 Cdc27 Cdc2l1 Cdc42 Cdc42se2 Cdk2ap2 Cdkal1 Cebpz Cggbp1 Chfr Chmp1b Chmp2b Chmp5 Chordc1 Chst14  45 20 43 48 14 83 18 21 852 18 17 37 25 48 11 28 392 105 47 22 280 18 46 12 56 53 17 10 39 50 28 15  12 7 23 8 2 37 0 12 204 12 8 34 10 22 4 11 248 8 12 17 126 8 23 1 22 47 11 2 4 33 9 4  7 8 27 10 4 44 8 6 284 7 6 21 14 13 4 8 165 24 7 7 106 3 18 2 18 30 3 1 17 13 2 3  28 16 15 11 17 62 7 12 399 26 7 13 11 33 8 14 122 71 29 11 181 9 19 6 20 53 8 8 54 33 9 5  23 4 27 7 9 21 3 7 108 15 11 20 8 14 7 5 98 37 12 15 81 5 17 4 17 40 5 7 25 24 6 3  Cisd2 Clcn4-2 Cldnd1 Clp1 Clptm1 Cno Cnpy2 Coasy Coil Colec12 Commd10 Commd2 Copa Cope Copg2 Cops2 Cops5 Cops7a Cops8 Coq5 Cox18 Cox5a Creb3 Creld2 Cript Crnkl1 Csrnp2 Cstf3 Ctnna1 Cul5 Cux1 Cyb5r3  90 12 31 21 70 24 111 13 23 11 42 100 68 143 12 133 72 14 62 25 11 819 12 95 13 15 12 20 56 18 13 33  55 0 10 9 45 3 40 6 10 2 15 61 12 26 1 25 29 3 8 5 3 422 7 15 3 2 3 16 25 14 8 15  47 4 8 14 56 4 61 7 10 3 22 30 22 35 3 29 21 7 3 10 4 664 9 18 5 1 0 6 21 8 5 22  48 14 24 7 29 6 43 10 10 15 24 50 41 88 7 64 36 4 23 12 3 503 12 17 6 4 2 12 23 8 6 25  39 10 12 10 26 7 27 10 4 6 9 22 64 31 4 18 23 4 18 6 3 394 8 13 4 3 4 18 24 5 7 17  163  Cyb5r4 D11Wsu99e D14Ertd500e D16Ertd472e D1Bwg0212e D2Wsu81e D4Wsu53e D930014E17Rik Dad1 Dctn1 Dctn4 Dctpp1 Ddah1 Ddb1 Ddt Ddx51 Decr2 Dem1 Dhfr Dhx15 Dirc2 Dlat Dlc1 Dld Dlg3 Dna2 Dnajb11 Dnajc7 Dnalc4 Dock1 Dolk Dpm3  26 23 33 13 33 28 94 9 144 12 37 64 212 39 229 13 12 16 122 207 29 52 36 27 50 17 15 57 15 46 14 93  16 8 11 0 7 18 61 2 85 6 18 30 54 7 42 1 9 6 80 108 18 17 16 9 25 3 7 14 7 18 8 28  22 9 12 5 4 14 31 4 97 5 13 27 80 13 40 2 9 3 77 96 10 20 11 11 26 5 5 20 5 22 5 37  17 12 10 9 8 10 180 7 113 3 26 7 132 23 119 4 16 3 49 109 20 21 29 18 27 6 4 16 5 32 9 25  15 6 5 8 8 6 94 11 60 3 13 25 59 26 48 2 14 3 63 46 29 15 13 5 31 2 5 22 2 13 2 74  Dpp9 Dr1 Dvl2 Dynll2 Dynlrb1 E2f6 Eapp Ebag9 Echs1 Eefsec Efr3a EG621842 EG625054 Eif3i Eif4b Eif4enif1 Elf2 Elp3 Elp4 Endog Epha2 Ercc3 Eri1 Erlin1 Erlin2 Exosc4 Exosc5 Exosc7 F11r F8a Fahd2a Fam108c  23 31 34 87 116 135 20 12 28 27 34 17 13 110 20 65 15 56 14 18 11 16 34 49 44 55 20 98 62 11 9 45  4 6 13 44 44 80 3 2 11 8 21 5 12 63 17 45 0 27 2 10 5 7 23 27 25 26 7 34 43 0 6 11  14 5 13 61 49 89 5 2 18 9 15 1 3 43 14 27 0 37 6 6 5 9 22 17 21 40 13 16 45 3 5 5  18 13 19 43 48 58 15 5 35 21 36 3 5 30 6 24 6 23 4 12 8 5 16 42 27 18 5 20 66 6 8 18  11 4 6 41 52 35 6 4 27 11 27 3 10 22 10 29 2 12 9 9 10 6 14 32 17 32 4 23 38 12 10 13  164  Fam114a2 Fam122a Fam128b Fam168b Fam3a Fam54b Fam73b Fam92a Fam98b Fat1 Fbxl11 Fbxo30 Fgfr1op Fgfr1op2 Fkbp5 Fndc4 Foxk1 Frg1 Frs2 Ftsj3 Fubp3 Fus Gabpb1 Gak Gas1 Gatad1 Gbl Gde1 Gemin7 Gfpt1 Ghitm Gins4  99 17 14 33 18 59 12 78 38 36 61 14 23 15 36 31 22 18 9 55 172 348 20 27 16 51 10 74 73 114 200 17  49 2 3 15 10 37 5 43 25 6 12 12 16 8 14 13 3 3 2 23 53 181 5 10 0 18 8 26 20 91 76 7  27 5 3 11 5 50 5 22 20 7 14 7 13 8 11 11 12 8 3 10 69 189 5 11 0 24 5 38 44 43 87 12  70 11 5 15 15 41 8 29 11 23 29 10 12 10 15 7 14 10 6 10 92 138 8 28 11 18 2 35 42 71 146 5  20 4 6 14 9 30 3 32 23 37 10 7 14 12 15 20 12 8 11 8 34 115 4 11 13 20 4 16 22 29 84 8  Gldc Glrx5 Gmcl1 Gnal1 Gnpat Gnpda1 Gnptab Golim4 Golph3 Gopc Gosr1 Gphn Grpel1 Gtf2e1 Gtf2e2 Gtf2h4 Gtf3c1 Gtl3 Gtpbp4 Gulo Gzf1 Hadh Hars Hars2 Hcfc1 Hhex Higd1a Hint1 Hnrnpa2b1 Hnrnpu Hnrpll Hsbp1  22 39 34 17 79 15 16 43 66 22 12 21 116 57 38 69 150 77 129 138 22 94 45 11 51 168 19 118 1574 682 136 132  0 13 26 11 7 4 7 21 17 18 7 4 21 41 8 50 120 20 36 30 13 46 22 7 15 47 10 34 1012 298 56 38  13 23 14 7 31 5 8 28 30 12 5 11 61 14 10 26 78 13 21 33 10 55 17 4 10 89 12 48 719 230 80 23  14 18 16 4 31 18 10 50 51 19 11 24 39 24 17 23 94 31 44 23 24 104 14 11 24 60 11 62 672 239 73 59  3 20 12 7 13 19 6 49 30 17 7 15 42 12 4 15 107 10 34 41 12 60 13 7 12 85 8 36 443 123 29 24  165  Hsd17b10 Hsd17b12 Hsd17b4 Hspa14 Hspb11 Hsph1 Huwe1 Ift20 Igbp1 Ilf2 Ing3 Ino80c Ints8 Ints9 Ip6k1 Ipo8 Ipo9 Ipp Irak1 Isg20l2 Itfg2 Itgb1bp1 Itih2 Jak1 Jarid1b Jtv1 Jub Kank1 Kcmf1 Khdrbs3 Kin Klhl2  62 190 25 11 25 280 86 38 60 61 28 28 33 17 10 84 158 14 52 34 10 40 600 71 57 75 17 23 47 20 12 20  11 65 9 10 3 78 26 6 15 15 11 7 6 2 3 41 72 4 17 9 3 5 533 16 11 24 3 5 8 11 11 7  20 172 10 4 4 119 34 7 20 18 3 8 5 2 2 29 54 6 12 7 4 14 331 26 4 25 6 3 21 7 2 12  48 192 23 6 10 55 37 27 27 18 12 9 10 5 5 43 97 8 40 7 8 13 369 68 16 21 7 22 27 5 5 9  33 168 26 3 5 48 62 13 13 8 8 7 4 3 5 38 23 2 10 14 2 11 306 29 15 28 7 7 25 22 6 4  Klhl21 Klhl5 Klhl9 Kpna1 Ktelc1 Kti12 L3mbtl2 l7Rn6 Lage3 Larp5 Lats1 Lemd2 Lgr4 Lias Limd1 Lin9 Llgl1 Lmf1 Lmo4 Lmtk2 LOC100044566 LOC100045439 LOC100046343 LOC100046483 LOC100047601 LOC100047839 LOC100048439 LOC676546 Lphn1 Lrrc57 Lsm1 Lsm14b  12 14 58 19 17 54 19 15 35 30 19 25 22 54 44 19 13 17 27 11 6 33 108 37 72 12 33 12 17 18 44 28  3 8 12 11 5 9 11 4 8 24 9 11 6 17 11 12 13 4 11 6 4 22 30 13 39 7 24 4 6 9 23 14  2 7 10 7 3 6 4 4 12 25 9 6 4 16 17 5 5 5 5 7 4 16 25 15 53 7 25 3 5 3 6 6  4 11 30 6 4 13 4 5 11 42 18 19 19 35 24 7 4 10 21 4 11 21 63 17 56 5 43 7 8 14 11 7  10 14 11 5 7 10 7 3 7 28 10 4 7 8 14 3 7 4 4 5 7 12 19 5 41 3 35 7 5 4 10 11  166  Lsr Lypla2 Lyrm5 Lztr1 Macf1 Man1b1 Manea Map1lc3b Map2k5 Map3k7 Mapre2 Mccc2 Mcl1 Mcm10 Mdh2 Mdm2 Me2 Med14 Med16 Med20 Med27 Med28 Mepce Metap1 Mex3c Mff Mgat2 Micall1 Mknk2 Mlec Mll1 Mll3  17 43 15 33 57 13 23 36 14 12 76 34 44 14 268 34 12 28 21 12 20 116 32 50 79 58 68 24 73 69 25 12  9 20 8 24 22 6 2 17 6 6 58 15 23 11 210 24 7 19 2 1 7 86 17 17 12 29 36 17 51 41 4 4  8 15 9 15 14 5 7 30 8 4 41 9 33 5 172 11 2 14 5 3 5 38 20 17 8 36 22 16 44 37 9 4  8 26 17 15 57 7 12 28 7 9 37 17 46 6 121 33 6 19 7 7 7 52 11 27 28 43 52 17 43 30 24 14  9 7 17 16 25 15 24 23 10 3 28 22 45 10 98 26 2 9 4 6 4 29 20 10 17 30 36 18 25 15 7 8  Mllt1 Mmgt1 Mphosph8 Mpst Mrpl14 Mrpl39 Mrps14 Mrps17 Mrps28 Mrps34 Mrps7 Msl3 Mtch2 Mtpap Mtrf1l Mttp Mylc2b Myo10 Myst1 N4bp2l2 Nadk Narg1 Nars2 Ncapd3 Ncbp2 Nck1 Ncoa3 Ncor2 Ndnl2 Ndufa9 Ndufaf2 Ndufb2  15 29 17 12 53 91 22 98 46 23 442 16 22 22 15 21 121 30 27 17 78 58 20 15 13 20 11 39 57 110 103 72  10 10 6 2 18 31 2 24 9 8 264 0 4 4 8 17 31 16 4 15 79 33 8 4 0 6 6 21 22 41 47 27  5 14 9 8 22 51 5 43 6 14 300 2 4 5 4 9 50 26 4 11 36 12 8 1 2 9 1 18 14 58 13 47  4 19 8 6 17 25 9 36 10 7 201 12 13 12 8 9 85 23 6 16 43 15 6 8 3 13 5 29 18 57 15 24  9 7 10 3 13 80 5 40 6 16 183 4 5 4 11 6 48 19 4 12 43 13 5 3 7 7 10 34 18 41 36 32  167  Ndufb7 Ndufb8 Neu1 Ngdn Ngrn Nif3l1 Nipsnap3a Nme3 Npepps Nr2c1 Nsbp1 Nsf Nsmaf Nsmce1 Nt5c Nt5c3 Nudt14 Nudt19 Nudt2 Nufip2 Numa1 Nup210 Onecut2 Ormdl1 Otud6b P2ry5 Paics Pak2 Patl1 Paxip1 Pcbp4 Pcgf6  105 97 15 54 19 23 17 26 44 15 43 34 23 31 37 29 22 52 14 47 38 30 69 64 39 10 218 16 18 22 55 31  32 17 8 21 12 14 2 11 23 1 19 25 14 25 15 8 6 36 7 16 11 12 18 13 33 6 148 7 3 15 12 6  61 55 10 19 7 10 5 6 12 3 13 15 7 11 16 8 10 29 7 16 7 19 16 10 5 3 107 7 7 5 16 5  38 41 13 17 10 7 9 8 24 9 25 19 15 8 7 24 7 30 5 35 14 7 40 24 14 17 92 11 11 6 4 7  51 35 16 13 5 4 9 4 16 3 6 10 18 11 12 19 27 27 5 6 10 12 15 49 16 8 43 3 9 12 16 6  Pcmt1 Pcmtd2 Pcnx Pde9a Pdgfc Pes1 Pex13 Pex19 Pfdn2 Pgp Phactr4 Phax Phb Phf2 Phf5a Phip Picalm Pik3c2a Pitpna Pldn Plscr3 Pno1 Poldip3 Pole4 Polr1c Polr1d Polr3e Pop4 Pop5 Pot1a Ppm1g Ppme1  41 20 12 17 10 17 23 19 128 80 60 15 45 18 92 42 51 25 28 24 19 93 35 39 185 57 38 61 72 10 179 23  13 16 11 12 0 6 6 10 67 39 11 8 15 13 51 22 32 11 12 7 7 48 10 7 79 31 5 19 20 0 102 6  18 16 5 7 0 3 5 11 88 43 22 5 15 10 54 11 16 24 12 6 7 42 18 10 58 37 10 8 41 2 76 8  33 31 8 8 1 3 16 10 24 32 43 4 13 17 26 26 72 44 19 7 14 27 19 18 20 21 9 18 21 8 38 13  13 19 6 6 3 3 20 10 52 14 14 10 17 7 27 19 34 23 12 6 12 18 7 20 52 21 6 22 14 3 62 5  168  Ppp2ca Ppp2r2d Ppp2r5e Ppp6c Ppt2 Pptc7 Prdx3 Prep Prkacb Prkar1a Prkcsh Prkrip1 Prmt6 Prnp Prpf8 Prr14 Prr3 Psip1 Psma1 Psmc3 Psmd12 Psmd8 Psme4 Psmg3 Ptcd2 Pten Ptges3 Ptpn1 Ptprf Ptprg Pum1 Pvrl2  338 28 22 34 8 30 365 93 13 145 27 22 17 48 136 23 98 33 191 88 70 75 23 32 44 20 73 50 498 14 37 28  107 11 15 13 4 25 261 23 6 68 14 5 6 1 79 6 36 16 103 46 21 52 7 6 22 6 32 10 255 0 6 19  89 19 12 12 5 12 229 32 6 48 19 9 4 2 98 3 26 5 54 49 20 43 7 12 23 6 14 30 216 6 9 5  103 17 15 28 10 40 274 25 11 107 17 5 6 82 38 11 35 12 52 27 17 33 12 3 15 19 19 19 115 9 21 4  141 10 15 16 9 22 265 20 8 42 22 4 3 33 47 9 18 6 135 21 32 26 28 10 18 26 23 10 165 5 15 7  Pycr2 Rab1 Rab10 Rab14 Rab1b Rab22a Rab2a Rab40c Rai14 Rassf1 Rassf3 Rassf6 Rbbp6 Rbm18 Rbm22 Rbm28 Rbm7 Rbx1 Rce1 Reck Rg9mtd1 Rmnd5b Rnf10 Rnf139 Rnf160 Rnf44 Rnf5 Rnmt Rpap2 Rpgr Rpp14 Rps19bp1  43 184 54 289 77 16 62 11 40 12 25 36 51 21 79 39 14 44 17 14 32 22 86 20 31 71 109 19 12 27 42 26  13 88 11 150 15 7 10 3 15 6 9 29 24 7 34 15 8 31 2 5 8 11 84 3 12 8 37 5 0 4 11 5  13 90 13 148 43 9 13 7 12 2 5 15 20 5 23 27 12 15 4 1 9 13 26 3 9 9 29 6 1 2 14 4  7 147 33 242 41 12 38 5 25 13 19 26 37 13 22 19 7 10 10 12 19 11 29 8 17 24 64 5 4 8 17 7  6 101 18 170 24 17 12 3 14 5 11 20 25 19 52 22 8 16 4 6 5 17 48 2 7 10 18 3 4 6 7 4  169  Rps6kb1 Rpusd4 Rreb1 Rrp9 Rtn4ip1 Rwdd4a Sae1 Samd4b Sap30 Sar1b Sars Sbds Scamp3 Scmh1 Sco1 Sdhd Sec16a Sec16b Sec62 Sema6a Sephs1 Setd5 Sf3b1 Sfmbt1 Sfrs2ip Sfxn1 Shb Shkbp1 Siah1a Sidt2 Sipa1l2 Skp1a  24 24 40 15 16 48 187 26 74 67 40 25 44 13 20 421 28 12 144 7 50 23 96 16 16 75 41 13 22 43 21 137  11 20 16 8 13 16 49 7 14 45 27 17 8 7 2 300 15 7 42 5 30 6 31 8 9 19 8 5 3 5 11 104  4 7 21 6 4 25 73 3 14 49 21 15 11 6 6 167 11 3 35 2 25 9 19 7 10 21 10 3 6 16 3 40  14 7 41 2 3 25 31 12 17 61 17 12 23 7 10 170 18 9 98 10 22 19 63 9 7 74 16 5 10 53 3 64  8 6 17 3 11 8 29 4 13 40 11 20 16 7 4 400 13 8 67 7 15 6 22 3 12 46 42 3 5 24 14 84  Slc16a1 Slc25a11 Slc25a17 Slc25a20 Slc25a5 Slc37a3 Slc3a2 Slc4a7 Smchd1 Smcr7l Smek1 Smg7 Smpd2 Snap23 Snap47 Snapc1 Snd1 Snrk Snrnp48 Snx18 Snx7 Socs6 Socs7 Sod2 Sos1 Spg7 Spr Sra1 Srpr Srrd Stam Stra13  117 80 67 52 503 18 10 12 23 10 21 90 13 26 54 19 147 19 11 20 18 23 12 14 18 13 36 56 50 20 21 43  25 32 29 28 178 3 3 0 5 9 8 47 4 16 20 0 60 7 0 3 5 7 2 6 7 11 6 18 11 2 15 34  18 35 16 44 320 4 4 0 3 2 8 12 3 10 19 2 78 8 1 2 5 7 3 5 12 6 20 21 10 5 8 16  56 44 47 54 178 11 8 6 6 3 18 59 9 22 26 4 110 16 10 12 8 9 3 8 8 6 16 34 38 7 11 21  14 64 41 38 173 13 4 2 4 3 13 60 5 11 29 4 42 5 6 5 7 7 3 9 5 8 28 19 31 3 15 12  170  Strap Strbp Stt3b Stx3 Stxbp6 Sucla2 Sumf1 Syap1 Synj2bp Taf1 Taf11 Taf12 Taf13 Taf15 Taldo1 Taok3 Tapt1 Tbc1d2b Tbca Tbp Tcea3 Tcf21 Terf2ip Tfam Tgds Tgfbr1 Thap11 Thoc3 Timm8b Tipin Tiprl Tjp1  239 34 127 12 42 87 21 23 29 13 41 12 11 19 174 41 24 24 76 26 59 10 10 204 16 16 45 66 433 37 13 35  160 5 93 2 25 26 18 3 9 3 8 4 4 9 69 15 14 17 16 13 14 0 2 69 11 0 16 15 132 6 3 10  167 13 89 3 9 54 12 7 9 7 15 2 2 2 85 34 7 18 30 17 30 0 4 122 11 2 23 17 90 7 5 7  95 25 106 2 25 61 27 18 19 8 19 6 4 21 105 21 22 28 18 16 54 10 8 82 15 7 9 15 170 8 6 11  91 20 94 7 12 39 22 20 29 8 13 4 8 10 24 24 16 20 41 14 25 12 3 44 10 4 7 13 93 5 3 19  Tlcd1 Tle1 Tm9sf3 Tmco7 Tmem123 Tmem128 Tmem14c Tmem167 Tmem167b Tmem184c Tmem188 Tmem216 Tmem57 Tmem93 Tmem98 Tnfaip1 Tollip Top1 Topors Trappc1 Trappc3 Trappc4 Trim32 Trip12 Trmt6 Trrap Trub2 Tsen2 Tsg101 Tsn Tspan12 Tspan14  27 25 61 10 51 41 128 126 10 26 27 16 29 107 23 60 19 165 20 40 21 43 18 115 26 61 29 16 90 289 39 22  12 2 28 2 28 14 82 59 8 13 13 7 14 40 6 21 8 113 11 11 7 11 3 60 6 33 16 11 24 63 11 10  23 7 14 1 11 11 94 58 5 11 14 3 7 45 12 27 17 51 5 13 4 6 4 31 12 23 7 6 41 71 2 8  13 14 45 3 36 15 89 56 8 22 19 4 14 46 11 34 24 110 8 25 5 13 3 70 7 30 5 3 36 108 15 17  14 11 14 4 42 12 89 30 5 18 16 2 8 17 10 13 15 37 19 8 19 9 3 42 12 37 9 5 33 48 30 4  171  Tsr2 Ttc13 Ttc33 Ttrap Tubgcp3 Tubgcp5 Tusc4 Txn2 Txndc1 Txndc13 Tyw1 Uaca Ube2b Ube2d2 Ube2k Ube2l3 Ube2r2 Ube2w Ube3c Ube4b Ubl3 Ubn1 Ubxn1 Uchl3 Ugcg Uhrf1bp1 Uhrf2 Unc119b Unc50 Unc84a Upf3b Uqcrfs1  24 33 15 57 13 14 14 142 60 13 25 11 133 31 84 93 197 20 54 117 28 19 60 44 41 46 41 52 36 19 19 501  14 9 3 4 8 5 3 39 28 0 9 3 50 15 41 40 53 4 33 37 4 14 28 25 28 21 34 27 18 8 8 159  7 9 5 3 3 3 6 59 18 2 8 5 38 21 31 45 40 7 28 30 8 8 28 12 21 16 15 16 13 4 5 200  8 10 7 20 7 8 11 62 23 11 6 5 107 12 26 30 115 12 18 61 31 15 49 5 36 26 22 12 12 11 13 234  8 11 4 12 9 7 7 27 12 4 4 6 103 17 17 23 26 6 15 14 21 14 24 8 21 15 8 20 18 4 8 217  Usp25 Usp8 Utp14a Utp18 Vapa Vhl Vps11 Vps16 Vps26b Vps36 Vps54 Vti1b Wbscr16 Wdr22 Wdr40b Wdr51b Wdsof1 Wnk1 Wrnip1 Xiap Xpot Xpr1 Yaf2 Yeats4 Yes1 Yipf4 Yme1l1 Ythdf2 Ythdf3 Ywhaz Zbtb1 Zbtb39  41 44 13 58 102 57 11 19 41 34 15 48 26 10 10 27 46 63 30 46 37 20 58 72 13 63 66 15 42 95 20 13  24 13 4 43 82 13 4 13 22 9 3 8 18 7 2 13 18 10 5 48 15 7 17 24 7 17 36 2 27 17 1 5  34 12 4 12 41 7 4 9 21 6 2 14 6 5 3 10 12 9 1 29 18 19 20 20 4 27 35 5 22 31 10 2  65 21 4 10 68 28 10 11 18 11 9 18 11 4 3 12 18 38 8 74 17 17 25 24 7 51 62 10 41 57 8 7  49 27 2 23 30 15 3 5 25 4 6 23 8 7 7 4 15 32 21 38 10 12 12 13 7 55 25 5 21 20 5 8  172  Zbtb41 Zc3h15 Zcchc8 Zdhhc5 Zfand6 Zfp148 Zfp318 Zfp574 Zfp598 Zfp740 Zfp787 Zfp830 Zfp91 Znhit1 Znrf2 Zranb2 Zrsr2 Zw10  11 25 33 46 37 30 15 10 41 36 18 17 65 12 20 69 26 11  4 8 12 33 23 6 8 2 6 13 12 5 34 0 2 19 13 7  7 5 7 26 10 6 6 1 9 11 6 8 25 5 3 8 9 4  11 12 27 16 38 20 12 2 24 19 10 12 35 5 11 43 10 5  6 9 7 37 43 13 18 2 8 14 5 4 59 2 4 17 10 4  173  Appendix C Cluster C genes and their expression in the five Tag-seq libraries (in tags per million).  Symbol 0610037L13Rik 1100001E04Rik 1110007A13Rik 1110007M04Rik 1110032E23Rik 1110036O03Rik 1500011H22Rik 1500012F01Rik 1700021F05Rik 1700029F09Rik 1700037H04Rik 1810043G02Rik 2010007H12Rik 2010309E21Rik 2210016L21Rik 2310008H09Rik 2310022M17Rik 2310061C15Rik 2310065K24Rik 2310079N02Rik 2410042D21Rik 2610024G14Rik 2610027L16Rik 2610036L11Rik 2610039C10Rik 2610528E23Rik 2700029M09Rik  E10.5 DLK1+ 42 11 52 78 22 34 38 183 20 49 17 11 22 70 27 26 45 44 45 30 15 15 43 47 66 21 57  E12.5 DLK1+ 9 2 25 40 3 6 9 32 3 28 3 7 12 26 15 9 2 23 16 18 9 6 37 29 22 4 12  E14.5 DLK1+ 24 2 18 53 5 12 9 65 6 12 4 6 12 19 14 7 14 19 17 4 3 8 9 20 16 7 15  E16.5 DLK1+ 30 2 30 35 2 12 7 85 7 13 6 4 7 19 13 5 18 11 27 11 8 4 11 15 20 6 13  adult liver 2 0 3 12 0 0 0 6 0 7 0 0 0 8 3 2 5 3 2 2 0 0 2 2 7 0 2  2810004N23Rik 2810417H13Rik 2900092E17Rik 3000004C01Rik 5730409G15Rik 5730470L24Rik 5930416I19Rik 6030405A18Rik 6230416J20Rik 6530401D17Rik 6720456B07Rik 9430015G10Rik 9830001H06Rik Aaas Abhd8 Adprh Agpat4 Alg2 Amz2 Anapc10 Ankrd49 Anp32b Anp32e Anpep Ap1s3 Aplp1 Aprt Arf3 Arhgap22 Arid2 Arid3a Arl2bp Arl3  35 211 32 14 23 19 45 15 13 75 183 14 12 35 20 41 11 37 22 12 11 13 186 44 17 20 124 35 16 53 14 30 37  24 89 22 4 8 13 35 9 3 35 78 9 4 24 7 15 5 13 8 5 4 5 87 18 12 16 51 10 12 22 9 11 10  8 58 16 6 9 6 8 5 4 51 60 6 4 13 3 9 0 13 4 5 5 0 102 24 4 9 30 7 11 6 9 12 4  11 71 16 7 11 10 23 4 4 22 104 10 6 10 3 16 2 19 14 6 5 3 54 25 4 5 53 18 6 31 1 7 10  2 0 4 0 2 3 2 0 0 0 13 0 0 0 0 2 0 0 0 0 0 0 5 5 2 0 7 0 0 0 0 2 3  174  Arl5a Armc6 Armcx2 Ascc1 Asf1a Aspm Atad2 Atp10a AW555464 B230118H07Rik B230208H17Rik Bat1a Bat3 Bax BC017612 BC023882 BC057552 Bcas2 Blmh Bmyc Bola1 Bola2 Bola3 Btbd2 Bub1b C030046I01Rik C1qbp C2cd3 C330027C09Rik Cacybp Cad Calm2 Calm3  14 64 67 20 45 14 119 14 59 28 23 197 89 37 24 25 34 240 60 58 19 124 55 25 68 18 1616 14 20 278 38 803 133  14 38 32 5 24 9 53 5 46 11 11 44 22 11 4 7 26 64 9 23 12 42 10 13 37 9 1211 8 9 154 16 180 32  6 46 30 9 23 4 47 4 24 9 9 41 30 8 6 2 18 64 18 7 12 24 12 15 36 2 393 4 3 148 16 295 52  11 28 17 11 12 12 72 0 38 14 12 56 35 6 10 14 19 129 16 24 6 25 19 9 33 7 371 6 12 44 6 481 54  0 4 5 0 2 0 11 0 2 0 3 13 7 2 0 0 0 21 5 0 0 3 4 4 0 0 55 2 0 27 0 20 5  Cand2 Casp2 Ccdc12 Ccm2 Ccnb1 Ccnb2 Ccnd1 Ccne1 Ccnf Cct2 Cct4 Cct6a Cct7 Cd276 Cdc16 Cdc25a Cdc73 Cdca4 Cdca5 Cdk7 Cdkn2aipnl Cenpa Cenpe Cenpf Cenph Cenpn Cfdp1 Chaf1a Chchd6 Chd1l Chek1 Chrac1 Cirh1a  100 11 44 26 108 73 363 208 46 866 264 300 276 20 38 33 12 58 19 90 23 158 23 21 56 70 98 61 42 17 20 28 205  56 2 12 16 50 47 65 86 18 542 68 97 64 7 15 14 6 11 6 10 12 89 13 9 39 60 19 24 25 5 11 8 133  18 3 12 7 29 30 97 87 27 462 28 135 76 5 16 3 2 36 12 22 8 51 7 7 15 39 24 20 30 4 1 10 105  11 2 13 17 24 44 48 31 23 234 80 97 81 5 20 5 8 17 6 32 10 64 17 17 23 30 50 28 17 6 8 15 56  0 0 4 0 0 0 20 0 0 64 8 10 27 0 2 0 0 5 0 0 0 3 0 0 0 2 4 0 2 2 0 0 17  175  Cited1 Ckap2 Ckap4 Clic6 Clint1 Clns1a Clspn Cnih4 Cnot10 Cnpy3 Cnpy4 Commd5 Cops3 Coq6 Coro1c Cpm Cpsf4 Cpsf6 Cs Csnk1g2 Csnk2a2 Ctnnbl1 Cul1 Cul2 Cul3 Cyb561d2 D10627 D15Wsu169e D230037D09Rik D2Ertd750e Dars2 Daxx Dbf4  204 55 99 14 200 128 13 41 11 22 22 21 30 39 84 209 21 81 63 23 18 12 59 208 88 30 12 12 17 37 22 23 74  76 18 46 10 88 34 5 12 5 10 12 10 9 16 44 97 6 38 16 4 4 4 38 53 40 15 5 2 5 29 6 10 66  97 14 43 0 82 41 5 21 3 8 17 8 18 19 33 82 8 32 16 16 4 4 13 89 20 21 2 3 7 21 7 7 32  33 31 54 0 153 21 5 16 5 8 11 8 12 15 31 42 9 32 19 10 7 4 24 130 57 13 6 4 12 17 6 13 24  0 0 4 0 14 9 0 2 0 0 0 3 4 2 2 0 0 2 7 2 0 0 7 20 10 2 0 0 0 0 2 0 6  Dcakd Dcp2 Dcun1d5 Ddx19a Ddx21 Ddx41 Ddx47 Ddx49 Ddx55 Depdc1b Depdc7 Dgkd Dgkz Dhdds Dhodh Dhx30 Dhx36 Dlgap5 Dnajc8 Dnmt3a Dnttip2 Dohh Donson Dtymk Dusp9 Dut E130012A19Rik E2f3 E2f5 E2f8 Ebna1bp2 Ecd Ect2  54 24 124 91 97 19 42 26 11 32 28 20 21 19 21 20 64 11 62 43 45 73 27 77 490 289 21 30 61 11 64 106 56  11 12 27 16 40 3 14 19 11 27 7 8 8 14 5 12 21 9 13 8 13 55 15 12 206 152 6 7 21 3 22 20 37  13 11 36 29 4 7 26 14 2 18 8 4 6 13 6 5 13 4 28 11 13 36 9 14 214 37 6 2 9 2 3 32 42  13 7 63 43 15 7 11 12 4 21 15 15 14 8 6 5 19 7 23 11 17 34 17 15 62 54 1 5 13 3 14 37 35  2 4 9 3 9 2 2 3 0 0 0 2 0 2 0 3 5 0 3 0 2 7 0 2 0 2 0 2 5 0 6 12 0  176  Edc4 Eef1e1 Eef2 Efna4 Egln1 Eif1b Eif2b5 Eif2s2 Eif3d Eif3f Eif3l Eif4a3 Eif4h Eif6 Elac2 Elavl1 Emid1 Enoph1 ENSMUSG0000 0074747 Erp29 Exo1 Exosc3 Exosc6 F630043A04Rik F730047E07Rik Fam117b Fam178a Fancm Farsb Fbln1 Fbxo28 Fbxo45  11 36 50 11 151 138 40 105 110 291 690 158 334 204 18 355 22 48  6 21 6 0 46 61 15 38 39 59 230 60 59 43 9 76 15 8  5 5 13 3 36 75 22 44 24 66 327 63 54 113 11 63 10 11  9 4 14 1 83 85 11 39 27 91 243 75 141 129 6 123 3 10  0 3 3 0 7 15 4 5 10 13 80 8 16 7 0 40 2 5  202 32 19 62 97 20 16 14 34 16 71 74 31 12  58 4 10 32 33 11 8 6 15 7 32 17 10 2  67 12 9 22 17 8 4 0 7 5 25 11 15 3  105 15 2 25 25 6 9 5 19 9 28 15 19 5  21 2 0 2 0 0 0 0 2 0 6 0 3 0  Fbxw8 Fh1 Fis1 Fkbp4 Foxm1 Fscn1 Fst Ftsjd1 Fundc1 Fusip1 Fzd7 G3bp1 G6pc3 Gadd45gip1 Galnt10 Garnl4 Gars Gart Gas5 Gata3 Gbas Gga1 Gins2 Glis2 Glrx3 Glt8d3 Gmnn Gnb1 Gnl2 Golt1b Gpn1 Gpn3 Grb10  17 30 165 370 10 286 110 12 31 60 149 139 24 75 11 10 64 60 83 19 127 31 82 54 148 13 47 63 25 56 68 15 1104  4 11 91 155 8 236 63 0 8 14 31 27 9 12 11 9 16 32 30 10 56 9 47 34 26 6 16 42 4 40 33 12 342  6 8 68 209 5 84 39 4 7 11 24 19 18 19 5 5 20 19 30 2 39 15 18 34 90 3 10 22 6 31 23 4 259  10 12 81 78 5 76 66 3 12 19 35 34 5 31 3 2 15 12 60 10 72 19 26 26 62 9 8 29 11 26 17 6 609  0 4 16 38 0 3 2 0 2 5 7 4 2 5 0 0 4 7 2 0 10 3 0 0 18 0 2 8 0 4 2 0 0  177  Grinl1a Grwd1 Gtdc1 Gtf2f2 Gtf2h2 Gtf2h3 Gtf3c5 Gtpbp1 Gtse1 Gys1 H1fx Hat1 Hba-x Hbb-bh1 Hdgf Hdgfrp2 Heatr1 Heatr3 Hectd1 Hic2 Hirip3 Hmgn1 Hn1 Hnf1a Hnrnpa0 Hnrpdl Hs2st1 Hsp90aa1 Hspa4 Hspbp1 Hspg2 Hyou1 Ict1  388 57 14 15 27 184 39 20 29 11 72 87 83 109 318 50 45 16 54 26 17 1471 32 15 945 87 23 26 192 68 27 160 37  116 21 6 11 10 141 32 13 14 2 17 28 47 108 54 37 8 4 38 16 13 853 19 5 373 26 10 5 99 14 6 40 10  130 7 7 11 6 62 17 10 12 4 25 12 7 2 80 23 14 3 15 11 9 799 27 10 201 14 5 5 114 35 9 26 16  126 6 8 4 13 62 21 8 8 6 16 22 14 3 124 22 10 4 24 19 3 557 15 7 345 39 11 6 55 39 12 39 15  33 3 2 0 0 25 5 4 0 0 0 3 0 0 8 2 5 0 8 0 0 135 0 0 58 0 2 0 17 4 2 3 2  Ift122 Igf2bp1 Igf2bp2 Igf2bp3 Ints1 Ints2 Ipo5 Irf2bp1 Isy1 Isyna1 Itih5 Itpka Jmjd6 Kat2a Kbtbd7 Kctd15 Kctd5 Kif11 Kif20b Klhdc3 Kntc1 Kpnb1 Krr1 Lass5 Lbr Lcmt1 Lgr5 Limd2 Lman2l Lmnb1 Lmnb2 LOC100045677 LOC100047506  17 140 246 42 41 31 243 18 47 216 54 11 58 34 17 20 32 72 22 45 17 103 58 27 470 14 85 12 16 78 77 25 52  7 65 112 14 15 6 76 7 21 95 21 3 27 24 9 7 8 48 18 38 10 29 24 6 294 3 40 11 5 19 33 5 52  8 50 58 26 20 6 77 8 18 44 14 9 26 15 7 5 3 19 3 17 4 22 21 10 119 7 34 2 11 8 22 7 21  8 39 69 24 12 6 57 5 14 21 12 4 34 5 8 4 5 45 10 21 5 23 32 7 214 6 9 8 10 15 28 7 23  0 0 0 0 0 0 19 0 3 0 0 0 8 5 2 0 0 0 0 4 0 0 3 0 14 0 9 0 0 0 3 0 3  178  Lrfn4 Lrrc16a Lrrc40 Lrrc42 Lrrc59 Lsm11 Lsm12 Lsm3 Lsm4 Lsm6 Lsm8 Lysmd2 Mageh1 Map2k1 Mapk8ip1 Mapre1 Marcks Mars2 Mat2a Mbd3 Mbtd1 Mcm2 Mcm4 Mcm5 Mcm7 Mdm1 Med24 Melk Metap2 Metrn Mett10d Mex3a Mfsd10  11 12 38 31 17 13 42 24 163 88 50 13 19 148 16 32 829 14 73 23 17 75 118 139 419 17 31 31 16 32 50 34 16  3 4 20 6 4 7 21 2 35 35 16 3 5 26 8 11 208 3 65 9 6 36 27 25 209 4 12 16 2 6 9 9 4  3 1 12 4 3 8 11 6 58 46 7 2 9 24 3 6 122 5 19 9 1 42 25 37 176 6 21 10 4 4 11 3 4  2 4 17 11 3 4 16 2 50 34 10 2 12 59 4 15 395 3 25 14 8 21 32 23 110 8 20 12 7 1 9 5 7  0 0 0 3 0 0 6 0 2 5 2 0 0 15 0 2 20 0 8 0 0 2 2 0 2 0 0 0 0 2 2 0 0  Mfsd7b Mlh1 Mllt11 Mmachc Mmadhc Mmd2 Mnat1 Mphosph10 Mpi Mrpl11 Mrpl13 Mrpl16 Mrpl19 Mrpl3 Mrpl32 Mrpl34 Mrpl38 Mrpl4 Mrpl41 Mrpl44 Mrpl46 Mrpl49 Mrps18c Mrps2 Mrps22 Mrps26 Mrps30 Msh6 Msto1 Mta2 Mtap1s Mtf2 Mthfd2  13 42 13 26 52 44 20 20 21 64 188 83 16 28 88 84 31 56 21 248 37 21 51 42 47 11 46 41 12 91 26 35 60  9 18 4 6 22 13 4 5 6 34 106 32 12 5 18 8 19 28 5 95 9 10 9 32 14 4 6 15 7 14 12 17 27  5 17 1 5 24 28 5 2 8 34 115 51 5 6 25 21 21 15 6 105 21 15 12 20 33 1 14 9 3 20 12 19 14  9 20 5 6 32 22 8 3 5 28 98 34 5 9 48 15 13 28 7 108 16 8 18 14 15 5 17 5 2 30 10 27 10  0 3 0 2 4 0 0 0 0 10 24 10 0 2 6 4 3 0 0 18 4 0 3 7 7 0 4 5 0 9 2 0 0  179  Mtmr9 Myc Mycn N6amt1 Nacc1 Nans Nap1l4 Narg1l Nat12 Nat13 Ncapd2 Ncaph Ncdn Ncl Ndufab1 Ndufaf1 Ndufs8 Nedd4 Neurl4 Nhp2 Nhsl1 Nip7 Nmd3 Nme4 Noc3l Nol10 Nol11 Nol6 Nol7 Nol8 Nop10 Nop16 nop5  17 29 44 11 64 76 70 10 42 219 37 51 30 129 11 57 88 233 19 797 13 87 59 33 10 23 19 29 139 17 1356 38 449  10 10 32 3 29 34 16 6 17 83 15 30 12 69 6 17 29 132 9 358 7 23 19 18 8 5 15 9 74 7 970 19 200  9 3 28 3 37 41 20 0 9 58 10 18 10 41 6 12 33 58 2 215 1 22 23 6 1 7 7 8 90 4 777 18 184  10 13 9 2 22 38 29 7 28 104 10 25 16 16 6 32 41 102 6 103 5 16 24 10 3 8 6 10 64 8 330 10 149  0 0 0 0 5 2 3 0 2 23 2 0 0 4 0 4 10 31 0 67 2 2 4 0 0 0 2 0 21 0 60 6 25  Nras Nrf1 Nsfl1c Nsun2 Nsun5 Nt5c3l Nubp1 Nucks1 Nudcd2 Nup107 Nup160 Nup188 Nup62 Nup85 Nutf2 Odz3 Orc2l Orc4l Orc6l Oxa1l Pacs1 Paf1 Pafah1b2 Pafah1b3 Pak1ip1 Pak4 Pank2 Pank4 Pawr Pbx2 Pcm1 Pcolce2 Pcsk9  27 21 97 15 20 39 29 41 54 30 44 33 57 107 33 29 19 19 49 40 12 56 80 40 69 22 16 11 13 135 12 111 103  19 8 37 9 4 14 6 7 21 19 40 11 15 51 20 9 10 12 17 4 4 22 46 20 23 5 7 4 8 40 8 65 24  8 8 51 7 7 9 6 10 11 6 21 8 44 19 5 14 5 9 18 18 6 5 31 24 18 8 4 5 1 36 4 32 40  15 13 54 4 7 11 8 19 17 8 25 12 25 22 6 15 8 12 11 22 5 18 31 13 31 5 7 8 8 34 7 23 59  0 2 3 0 0 0 3 0 3 0 0 0 5 5 0 0 0 0 0 3 0 0 11 0 6 2 2 0 0 0 0 3 5  180  Pdap1 Pdcd11 Pdcd2 Pdia3 Pdia6 Pds5b Pdzd11 Pdzd8 Pelp1 Pfkfb3 Pfn2 Pgam5 Pgd Phf12 Phf23 Phpt1 Phrf1 Pi4kb Pigu Pih1d1 Plagl2 Plk4 Pole Polr1e Polr2i Polr3k Pop7 Ppig Ppp1ca Ppp1r8 Ppp2r2a Ppp4c Ppp4r4  12 21 24 11 381 32 56 18 41 88 62 164 51 23 50 21 23 13 19 73 30 31 21 26 77 65 40 33 11 38 26 126 23  1 14 2 10 228 18 44 5 20 42 16 88 16 13 29 6 4 0 8 12 10 13 10 12 33 17 13 14 3 18 9 42 10  3 10 8 6 154 16 32 7 21 25 7 55 22 7 19 6 9 7 6 45 6 4 11 8 31 19 20 4 6 16 13 20 3  4 8 10 7 148 19 37 10 20 51 8 49 22 16 18 8 11 7 5 44 7 15 12 4 18 20 17 16 6 7 12 68 5  0 0 0 0 45 0 2 0 2 2 2 15 4 3 7 2 2 0 2 6 3 0 0 0 10 7 2 2 0 5 0 2 3  Prdm4 Prkaca Prkar2b Prkcbp1 Prkrir Prmt3 Prpf31 Prpf4b Psma6 Psma7 Psmb2 Psmc3ip Psmd1 Psme3 Pspc1 Ptges2 Ptgfrn Ptk2 Ptov1 Ptprs Pum2 Pwp1 Pycrl Qsox2 Qtrtd1 Rab34 Rabepk Rabl3 Racgap1 Rad21 Rbm14 Rbm17 Rbm34  17 72 11 13 18 55 21 33 346 109 165 46 243 188 35 20 105 16 127 28 41 37 79 13 12 43 30 17 60 88 12 68 96  5 12 5 2 2 26 3 16 126 18 42 9 49 41 11 7 47 5 64 27 19 12 37 12 4 23 13 14 60 22 1 29 26  4 14 5 4 5 13 8 17 115 20 35 20 65 67 6 13 14 3 69 6 12 4 44 3 0 19 7 9 28 30 4 35 41  5 27 8 6 6 32 5 22 116 20 44 11 90 69 14 12 11 6 42 21 34 9 57 9 7 22 14 11 41 46 2 26 57  0 6 0 0 0 6 0 3 43 6 19 0 17 14 0 2 4 0 11 2 0 0 2 0 0 0 0 0 0 3 0 4 9  181  Rbm4 Rbm42 Rbm45 Rbmxrt Rcc1 Rcn1 Rfc3 Rfc4 Rfc5 Rfxap Rhobtb3 Rif1 Rnaseh2b Rnaseh2c Rnf2 Rnf219 Rnmtl1 Rpa2 Rpa3 Rpap1 Rpl14 Rpl31 Rpp21 Rps3 Rps5 Rps6 Rps6ka1 Rrm1 Rrm2 Rsrc1 Rtcd1 Rtf1 Ruvbl2  29 25 28 32 17 72 72 45 69 17 48 37 29 213 55 16 10 46 41 20 65 63 29 102 3175 186 21 102 313 25 92 73 192  11 14 20 11 3 21 15 9 31 0 13 27 18 141 29 5 6 10 11 9 49 54 15 55 787 116 12 26 223 8 13 32 146  14 10 15 9 6 9 37 6 26 6 3 14 9 88 13 4 6 17 2 7 37 25 8 64 584 108 7 19 113 9 31 27 99  13 10 15 7 2 32 22 11 16 5 17 28 10 78 42 6 5 19 8 5 15 12 11 20 723 68 5 28 114 9 22 35 40  4 3 3 2 0 4 3 0 8 0 0 2 4 33 5 0 0 0 0 0 2 3 3 5 290 28 3 2 5 3 5 10 20  Saal1 Scrib Scye1 Sdad1 Sec11a Seh1l Senp1 Serpine2 Sf3a2 Sf3b3 Sf3b4 Sfrs12ip1 Sfrs15 Sfrs2 Sfrs3 Sfrs6 Slc24a5 Slc25a33 Slc25a4 Slc2a1 Slc2a3 Slc38a1 Slc39a6 Slc4a2 Slc7a6 Slmo2 Smarca4 Smarcad1 Smarcd1 Smarce1 Smc3 Smn1 Smyd2  33 18 54 53 63 109 19 11 49 23 608 27 31 31 15 263 30 12 276 238 55 10 23 36 12 129 97 18 20 41 81 72 24  10 4 32 38 28 35 8 8 13 4 316 8 7 14 6 146 3 10 91 135 32 4 5 24 4 50 24 9 6 9 14 17 14  14 5 27 26 32 12 7 2 19 6 213 16 6 4 4 143 13 7 73 73 16 1 7 24 3 83 63 1 4 7 16 38 10  20 7 29 19 41 48 6 3 16 4 184 16 8 11 8 130 16 9 86 77 5 1 5 19 3 53 45 8 6 8 22 21 8  0 2 5 0 8 11 0 0 0 0 29 2 4 3 0 21 0 0 3 12 0 0 0 4 0 9 0 0 0 2 7 0 2  182  Smyd5 Snhg1 Snip1 Snrnp25 Snrnp27 Snrnp70 Snrpb Snrpd1 Snx2 Snx27 Sort1 Sox12 Spats2 Spc24 Spc25 Spcs2 Srp14 Ssbp4 Sssca1 Stam2 Steap1 Stt3a Sub1 Surf2 Suv39h2 Syce2 Syncrip Tacc3 Taf5 Taf6 Tbc1d19 Tbc1d7 Tbl2  44 147 27 31 21 189 374 280 35 20 26 85 15 29 59 211 145 13 35 28 17 67 36 24 38 23 111 33 22 48 12 15 13  19 100 17 6 16 54 121 268 27 8 22 25 4 6 23 51 37 9 7 19 9 24 10 8 12 13 64 27 17 12 4 3 8  36 45 17 11 5 54 112 28 4 2 13 19 4 12 29 65 54 8 9 7 8 23 5 4 10 14 61 14 7 7 1 11 8  23 54 12 5 8 95 104 35 28 15 8 22 2 8 27 71 94 2 6 16 3 46 10 11 5 8 56 20 9 22 6 5 5  0 12 4 0 4 8 31 9 2 2 0 6 0 2 0 22 13 0 3 2 0 2 2 3 0 0 7 0 0 2 0 0 0  Tceb1 Tcerg1 Tcfe2a Tex9 Th1l Thap7 Thoc1 Thop1 Thumpd3 Tia1 Timm10 Timm17a Tinf2 Tle6 Tln2 Tmco1 Tmed9 Tmem101 Tmem110 Tmem158 Tmem161a Tmem165 Tmem168 Tmem201 Tmem209 Tmem39b Tmem41a Tmem43 Tmem55a Tmem60 Tnpo3 Toe1 Tomm40l  24 74 83 13 16 10 18 72 25 45 14 46 21 22 37 18 90 29 20 66 20 49 18 17 70 12 49 20 121 11 126 11 47  6 25 16 3 6 2 13 38 3 38 2 11 17 11 17 5 27 7 9 36 8 45 10 10 56 4 16 10 25 5 23 4 19  7 22 29 2 3 6 5 28 5 13 5 7 5 16 16 6 17 6 18 34 6 18 6 6 25 5 22 7 23 3 37 3 20  12 41 23 7 6 5 11 5 10 37 4 10 10 3 14 10 21 7 8 8 13 29 13 11 29 3 10 15 53 8 32 4 21  2 4 0 0 0 0 2 3 2 2 0 2 3 0 0 0 0 0 0 0 0 4 0 0 6 0 3 0 7 0 7 0 0  183  Top2a Topbp1 Trim27 Trim35 Trim44 Trip10 Trip13 Trp53bp1 Tspan18 Tspan3 Tsr1 Ttc1 Ttc3 Ttc30a1 Ttf2 Ttll12 Tubb5 Ubap2 Ube2n Ube2z Ubfd1 Ubiad1 Ubr7 Ubtd2 Uchl5 Urb2 Urm1 Usp1 Usp21 Usp22 Usp39 Usp46 Usp7  149 13 138 51 53 14 86 23 12 214 86 26 177 18 25 11 455 34 31 45 15 29 25 29 15 11 36 34 37 44 69 11 49  28 6 61 15 26 6 43 15 3 82 35 12 48 6 15 2 171 9 15 19 7 7 11 11 9 4 10 14 19 10 17 6 22  30 5 76 18 17 6 44 3 2 28 46 13 89 10 14 4 168 7 7 27 4 10 9 9 3 2 26 11 11 7 12 4 16  67 5 35 18 21 5 20 12 3 68 38 16 117 7 9 3 107 11 8 25 4 8 11 8 5 3 20 23 26 21 14 3 32  0 0 21 0 4 2 2 0 0 2 7 4 7 2 2 0 7 4 2 6 2 0 2 0 0 0 0 2 0 3 2 0 6  Utp15 Utp3 Utp6 Vdac2 Vps25 Wasf1 Wbp5 Wdhd1 Wdr12 Wdr21 Wdr3 Wdr55 Wdr68 Wdr70 Wdr77 Wrb Xpo6 Yars2 Ywhab Ywhag Yy1 Zc4h2 Zcchc10 Zcchc17 Zcrb1 Zfp251 Zfp280b Zfp281 Zfp326 Zfp354c Zfp362 Zfp647 Zfp697  21 65 12 361 69 12 11 34 39 15 43 32 10 12 185 37 31 65 36 258 11 28 23 17 108 11 36 55 12 15 22 14 26  13 20 3 217 36 4 8 27 25 9 21 5 3 8 89 7 12 25 15 198 8 6 6 8 28 5 20 16 2 6 8 8 5  2 9 4 201 48 1 4 22 16 7 22 22 3 6 82 15 9 26 19 145 2 0 2 9 13 3 7 27 2 3 5 2 4  7 19 2 222 36 2 8 12 15 11 23 12 4 4 61 8 12 22 19 132 6 10 5 7 41 7 11 28 4 3 9 3 3  0 5 0 15 4 0 0 0 2 0 2 2 0 0 14 0 4 4 0 38 0 3 0 0 6 0 2 2 0 0 2 0 0  184  Zfp706 Zfp770 Zik1 Znhit3 Znhit6  140 11 16 42 14  86 3 3 9 8  66 6 5 15 5  48 7 7 9 2  24 0 0 4 0  185  Appendix D Cluster D genes and their expression in the five Tag-seq libraries (in tags per million).  Symbol 1110004E09Rik 1110014N23Rik 1190002F15Rik 1500003O22Rik 1700029G01Rik 1810034K20Rik 2310003L22Rik 2310014H01Rik 2410091C18Rik 2510012J08Rik 2610002J02Rik 2700099C18Rik 2810021J22Rik 2810452K22Rik 3110082I17Rik 4632434I11Rik 4930427A07Rik 4933411K20Rik 6030458C11Rik 6720460F02Rik 9130401M01Rik A230051G13Rik Abcc5 Acap2 Acot7 Acyp1 Adam17  E10 DLK1+ 23 21 17 38 2 18 16 6 20 19 42 25 5 14 100 11 6 15 19 38 68 7 28 23 14 4 33  E12 DLK1+ 20 36 21 39 15 30 49 18 30 69 55 52 14 20 114 13 10 20 29 61 302 12 48 32 22 15 40  E14 DLK1+ 8 23 10 13 0 27 21 10 17 35 17 29 4 17 67 5 6 16 10 53 99 9 37 16 13 5 27  E16 DLK1+ 6 15 14 17 5 11 9 12 10 6 14 44 7 4 30 4 7 11 19 12 16 4 32 26 10 4 23  adult liver 5 7 2 3 0 5 5 3 4 11 16 3 0 5 9 2 0 3 10 0 39 0 0 2 3 3 7  Afg3l2 Agrn AI847670 AK220484 Akap10 Alg8 Ankfy1 Ankle1 Ankrd32 Anks3 Appl1 Arl2 Asb8 Asns Asxl1 Atpif1 Atr Atxn2l Bag2 Bahd1 BC027231 BC031781 Bcs1l Bex6 Bpnt1 Brca1 Brd4 Brd8 Brip1 Brp16 Brpf1 Bysl Cap1  26 142 19 14 30 65 12 5 15 9 6 23 6 86 17 333 8 35 81 17 10 21 6 13 9 19 97 58 3 21 10 65 21  64 156 23 17 54 72 17 14 30 88 16 36 13 198 22 800 16 70 162 19 13 43 11 42 11 25 124 91 11 18 15 59 25  27 93 11 7 15 60 10 7 21 33 5 32 5 138 13 400 9 31 61 3 7 23 3 7 4 18 56 57 1 8 3 22 6  22 84 12 11 33 21 13 8 18 23 9 10 6 112 15 189 11 40 27 9 5 21 3 7 3 8 61 69 1 1 7 10 15  6 15 6 0 6 4 2 0 3 4 2 8 4 0 0 19 0 7 0 6 0 10 0 0 3 0 27 5 0 5 0 17 6  186  Ccdc43 Ccdc99 Cdc42ep1 Cdc42se1 Cdkn3 Cdon Cenpj Cep152 Cep68 Cep72 Chek2 Chst12 Ckap5 Cltc Commd6 Cox11 Ctbs Ctcf Ctnnb1 Cul7 Cybasc3 Cyth3 D19Bwg1357e Dcbld1 Dcbld2 Dclre1a Dcps Ddx19b Ddx42 Deaf1 Degs1 Dhx38 Dicer1  41 20 10 19 5 5 8 7 7 6 12 8 43 99 14 49 7 40 144 31 4 6 154 17 52 9 194 9 41 34 83 69 20  97 25 55 22 27 17 20 26 12 15 21 11 119 160 15 77 33 42 193 49 25 15 530 32 55 20 271 13 76 85 152 140 34  20 11 25 12 13 10 12 5 4 6 10 9 54 50 9 46 15 24 40 42 12 7 276 20 39 10 199 8 53 42 67 72 14  20 5 15 7 7 4 9 15 1 9 10 3 35 122 7 37 15 22 62 11 18 12 74 9 26 8 119 6 32 27 52 53 19  4 0 2 5 0 2 4 4 2 0 0 0 15 13 4 8 6 6 45 3 0 0 32 2 0 3 20 0 14 15 50 7 4  Dkk3 Dlk1 Dph5 Dpy19l4 Dsn1 Dus2l Dusp11 E130309D02Rik E2f1 Edem3 Eif2a Eif2b1 Eif2b3 Eif3k Ell3 Elof1 Elovl6 Elp2 ENSMUSG00000070510 Erf Exog Ext2 Extl3 Fam175a Fbxo31 Fdx1l Fgfr4 Fignl1 Fkbp10 Fkrp Frat1 Fut11 Fxr2  9 6094 25 2 17 18 35 9 9 29 47 59 14 408 19 22 69 58 13 13 13 16 5 15 16 60 57 39 15 5 13 10 12  11 7545 76 13 19 70 103 14 29 38 112 74 14 778 49 139 150 83 27 26 18 21 15 18 35 125 103 46 21 14 21 25 15  8 5083 22 3 11 35 66 7 11 18 62 54 8 461 24 55 83 74 7 14 14 7 2 7 16 40 45 28 3 4 8 4 8  3 4169 15 3 15 13 31 3 6 21 25 25 5 175 25 13 89 34 7 6 5 11 6 15 13 27 59 12 5 2 2 8 5  0 0 7 0 0 13 26 2 7 13 35 9 2 94 0 21 28 12 0 6 2 4 0 0 10 41 17 0 0 3 3 2 4  187  Fzr1 Gatc Gcat Gins3 Gipc2 Gm826 Gspt2 Gusb Hbb-y Hdac3 Hdlbp Hgsnat Hisppd1 Hkdc1 Hmox1 Hsf1 Hsp90ab1 Hspa9 Htatsf1 Hus1 Idh3a Ift140 Igf2r Igsf1 Ilkap Impact Ipo11 Ipo4 Ippk Itsn1 Katnb1 Kdelr1 Kif12  19 28 77 12 24 16 11 233 520 38 185 22 51 59 22 73 13 354 20 22 138 24 89 91 152 70 71 82 15 22 8 93 11  24 101 227 26 34 58 15 679 931 57 429 120 73 304 42 79 13 417 32 30 160 47 112 352 208 124 68 68 17 39 12 283 15  21 41 129 5 20 20 6 378 198 39 276 30 57 147 27 49 5 255 19 17 65 26 66 138 95 71 33 38 6 20 7 138 11  8 14 50 4 12 5 5 81 95 11 220 51 48 56 35 45 4 244 24 7 48 19 18 89 132 62 16 15 9 21 4 89 8  2 25 35 0 0 4 0 37 0 14 115 23 6 0 4 9 2 46 7 6 13 10 2 0 36 32 17 17 0 6 2 74 0  Klb Lars Letmd1 LOC100048019 Lyar Maged1 Map2k2 Mare Mcm3ap Mcm6 Mdn1 Men1 Mesdc1 Mesdc2 Mettl10 Mettl11a Mki67ip Mobkl3 Mov10 Mpg Mrpl54 Mtbp Mtg1 Mycbp Nadsyn1 Napa Ncoa5 Ndufa8 Ndufb11 Nedd1 Nf2 Nfatc3 Nfs1  34 56 82 3 121 507 154 16 17 459 30 18 61 51 40 22 44 76 9 9 163 10 14 11 14 19 17 42 75 32 23 16 151  91 198 211 12 132 568 411 21 27 589 52 23 66 56 74 24 55 100 12 32 218 13 16 18 61 40 14 59 248 85 42 44 145  45 69 135 4 78 305 238 8 17 471 12 4 23 43 47 18 39 56 8 14 140 7 12 12 37 29 9 27 99 41 35 12 75  47 28 110 0 19 389 93 10 16 300 12 3 29 16 18 11 10 52 5 4 59 6 7 7 25 20 4 37 40 38 8 8 93  10 29 29 0 11 49 94 3 7 10 3 2 5 5 6 2 17 32 0 5 15 0 0 0 5 5 3 19 75 26 12 8 22  188  Nop56 Nrm Nup155 Nup214 Orc1l Osgep Pa2g4 Parp1 Parvb Pcid2 Pea15a Pelo Pfkl Phb2 Pisd Pkn3 Plekha1 Plod3 Pms2 Pnliprp2 Polr2e Polr2h Polr3h Pomt1 Ppil2 Prc1 Prim1 Prkar1b Prkdc Prpf6 Prr5 Psen1 Psmc4  184 35 120 11 8 23 509 192 24 16 17 24 107 211 17 7 24 60 27 19 154 69 18 5 23 39 38 13 16 40 40 5 172  236 49 124 11 11 31 812 233 109 22 30 32 111 238 68 12 34 134 28 23 152 144 21 11 61 55 51 16 31 56 58 11 673  129 25 67 3 1 12 261 149 56 5 15 19 52 144 25 5 12 66 20 21 111 30 15 6 36 29 37 6 24 22 36 5 241  44 8 53 6 4 22 114 51 5 5 13 16 19 94 15 5 23 35 13 8 51 20 7 2 18 26 20 2 15 13 30 4 60  4 3 21 3 0 0 63 55 19 3 8 9 38 7 10 0 3 22 3 0 12 3 0 3 4 0 0 0 8 12 10 2 135  Psmd3 Psmd5 Ptdss2 Rabgap1 Rabl4 Rae1 Ralbp1 Rara Raver1 Rbmx Rcn3 Reln Renbp Rexo1 Rfc2 Rnf168 Rngtt Rpa1 Rpl21 Rpl23 Rpp30 Rpp38 Rrp15 Rsf1 Rwdd2b Sac3d1 Sart3 Scamp5 Scarb1 Sdccag8 Sec13 Sec23ip Sgol2  42 13 17 15 37 31 43 4 54 22 43 128 10 4 104 15 26 29 13 72 8 27 91 10 13 28 48 11 92 8 75 39 11  64 21 32 33 194 53 64 14 57 42 49 237 64 11 208 24 68 31 14 104 11 33 106 18 31 30 70 14 354 34 178 53 15  33 7 25 6 65 40 34 4 39 5 9 129 37 4 121 16 30 22 8 77 5 19 55 6 17 22 48 10 191 12 127 31 9  12 6 8 16 15 6 47 5 24 26 19 90 35 4 46 9 17 11 4 34 3 9 36 11 8 5 24 6 204 12 47 28 10  13 5 8 8 43 16 19 3 8 0 4 4 3 0 6 4 9 7 0 2 2 3 6 0 4 3 19 4 55 10 14 2 0  189  Shcbp1 Sipa1l1 Slc12a4 Slc1a5 Slc22a8 Slc5a6 Smc4 Snhg3 Snhg6 Speg Srp68 Srprb Ssna1 Ssr2 Ssr3 Stx4a Tbcc Tcp11l1 Terf2 Thap1 Tln1 Tmem184b Tmem199 Tmem48 Tmem51 Tmem9 Tmod3 Tmpo Tnfaip8 Tpcn1 Trim25 Trip6 Trmt11  22 8 15 43 6 24 101 123 258 15 63 59 70 315 510 19 2 14 15 9 3 68 38 20 14 106 14 14 20 26 16 30 22  27 13 35 129 11 33 121 133 437 22 153 64 96 702 505 141 21 15 22 11 10 82 44 44 23 152 34 24 86 33 36 172 27  16 4 13 53 2 15 50 80 222 9 80 40 76 554 196 31 3 9 12 4 3 46 24 16 15 123 14 20 20 16 14 78 14  14 1 8 38 6 10 48 42 135 5 42 26 46 439 316 31 6 2 12 3 0 56 16 6 10 49 14 9 9 18 9 38 12  0 3 6 2 0 4 9 24 114 0 24 2 7 63 96 15 0 0 0 3 0 18 11 5 0 11 8 0 0 8 10 34 4  Trnau1ap Tspan13 Tssc1 Tug1 Txndc12 Txndc5 Txnl1 Unc45a Uvrag Vars Vasp Wasl Wdr13 Wdr46 Wdr82 Wfikkn1 Wnk4 Xrcc1 Xrcc5 Yipf5 Ywhae Ywhaq Zc3h7b Zdhhc3 Zfp189 Zfp384 Zfp397 Zfp408 Zfp518b Zfp532 Zfp623 Zfp628 Zfp687  30 18 17 13 30 139 358 10 11 67 17 18 10 27 18 4 14 43 45 28 1089 544 42 56 4 16 4 6 16 14 5 4 3  43 23 24 16 29 455 436 21 33 95 76 64 21 31 27 10 14 61 49 37 1085 627 142 58 26 48 27 17 17 19 22 28 16  20 19 15 8 16 203 216 13 17 32 30 23 12 19 19 2 8 39 25 16 518 175 50 24 10 14 11 9 14 11 7 11 5  10 10 10 10 19 127 268 6 21 35 40 22 16 11 11 0 5 32 18 26 540 213 12 24 5 17 8 11 6 10 5 2 3  10 2 7 0 4 122 76 3 9 0 4 15 3 6 7 0 0 5 4 8 192 4 17 13 6 8 4 0 0 0 3 4 2  190  Zfpl1 Zh2c2  10 10  12 19  6 8  9 6  2 0  191  Appendix E Cluster E genes and their expression in the five Tag-seq libraries (in tags per million).  Symbol 0610007L01Rik 0610010F05Rik 1110008J03Rik 1110008P14Rik 1110037F02Rik 1110038D17Rik 1300012G16Rik 1600021P15Rik 1700020C11Rik 1810026J23Rik 1810030N24Rik 1810031K17Rik 1810055E12Rik 2010107H07Rik 2010305A19Rik 2300009A05Rik 2310016M24Rik 2410129H14Rik 2510049I19Rik 2610001J05Rik 2610002M06Rik 2610528J11Rik 2700078K21Rik 2900064A13Rik 3110001I20Rik 3200002M19Rik 4931406C07Rik 4933403F05Rik 5033414K04Rik  E10 DLK1+ 27 14 19 12 25 35 17 28 39 14 21 7 9 23 15 11 66 11 32 89 6 4 8 48 15 44 23 39 44  E12 DLK1+ 38 19 26 7 23 167 37 60 31 36 25 19 8 24 24 2 118 14 27 96 25 10 6 27 21 37 98 28 52  E14 DLK1+ 40 16 16 9 13 188 21 35 31 8 18 9 8 12 24 10 110 11 40 55 22 10 11 56 10 41 71 22 28  E16 DLK1+ 19 12 9 5 10 40 11 18 12 7 10 13 7 12 8 9 31 14 24 39 13 2 9 44 13 15 81 28 46  adult liver 33 15 13 16 19 183 45 60 20 15 24 23 10 33 15 18 60 19 27 50 38 4 11 58 27 23 161 44 41  5730403B10Rik 5730455O13Rik 5730469M10Rik 6330409N04Rik 6720456H20Rik 9030624J02Rik 9430023L20Rik A1cf A430005L14Rik AA960436 Abat Abca2 Abcd3 Adal Adck1 Agfg1 Agmat Agpat6 Ahsa2 AI316807 Akap8l Akr1c12 Akr7a5 Alas1 Aldh2 Anapc4 Angptl4 Ankrd28 Anks1 Anks4b Ap3m1 Ap3s2 Apeh Arid4a Arl8b  17 57 47 7 8 25 4 12 25 20 125 14 126 9 14 41 49 113 121 13 15 30 36 25 74 23 38 22 10 5 82 36 20 6 80  33 78 71 14 11 23 7 13 31 23 189 25 51 10 19 43 49 220 172 12 20 90 17 150 13 12 103 55 11 13 135 36 37 8 97  16 85 69 12 18 23 4 10 26 25 290 13 107 6 15 63 55 124 95 12 26 18 48 180 38 28 136 47 6 6 135 22 34 5 87  24 51 65 3 11 14 1 11 10 13 212 7 185 8 8 30 46 48 89 9 26 33 22 26 47 18 58 28 6 7 78 11 15 10 79  15 45 71 23 16 14 10 7 15 18 321 12 178 9 10 28 37 97 130 13 39 36 24 293 93 15 177 29 8 12 158 21 34 12 66  192  Arpc3 Arsb As3mt Asb3 Atg2a Atg5 Atp1a1 Atp5d Atp5sl Atp6v0a1 Atrnl1 AU022252 AU022870 Aup1 B230339M05Rik Bace1 Bag3 Bcar1 Bcar3 Bcas3 Birc6 Blvra Bmp1 Brf1 Brpf3 Btaf1 Bxdc5 C330016O10Rik C8a Calr Capn2 Carhsp1 Carkd Ccbl1 Ccdc25  63 30 20 13 14 36 78 238 19 18 12 9 57 17 17 8 21 15 19 6 64 30 18 15 18 27 93 5 23 59 17 37 21 23 21  123 66 32 4 24 14 78 441 18 16 12 13 41 22 19 17 25 42 34 10 123 82 11 29 21 45 158 18 97 139 22 14 25 140 19  68 64 22 13 27 14 110 704 22 19 3 16 53 30 8 10 23 20 29 13 77 43 20 32 13 29 98 10 44 44 29 25 40 223 12  33 26 34 7 18 14 46 137 13 17 5 3 20 11 9 5 21 14 21 5 73 28 22 20 16 21 40 2 85 12 18 21 22 82 9  155 45 36 12 26 49 126 647 36 28 14 14 35 31 10 11 34 28 16 12 59 86 28 28 12 27 69 12 91 116 24 54 58 214 32  Ccdc58 Cct8 Cdc42ep4 Cds2 Cebpg Chchd3 Chchd5 Ciao1 Ciapin1 Clcc1 Clcn7 Cldn12 Clpp Cmc1 Cobra1 Cog7 Cog8 Commd8 Copb2 Copz2 Coro1b Cox4nb Cox6a1 Cox6c Cpn1 Cpsf1 Cradd Crebzf Crk Crtc1 Cry1 Cry2 Csnk2b Cstb Cstf1  88 387 23 12 34 102 10 36 43 29 12 63 82 59 21 17 59 22 82 11 22 15 38 251 80 57 3 27 96 7 27 7 30 141 17  93 791 19 11 42 55 17 62 72 15 13 68 144 61 24 21 38 27 63 44 34 23 65 234 92 84 8 38 108 14 27 23 52 503 27  65 476 21 12 22 51 22 70 13 18 9 100 76 58 18 16 46 11 65 25 37 13 33 379 117 58 5 22 130 9 21 22 43 145 38  28 108 26 9 20 46 7 24 7 18 12 46 15 46 10 12 20 17 54 22 17 4 25 164 74 25 5 37 62 4 18 7 9 90 12  71 410 22 21 42 144 13 55 35 38 9 55 140 48 18 37 37 18 48 24 27 17 35 566 74 50 13 30 72 5 46 13 61 200 26  193  Cstf2 Cstf2t Ctdp1 Ctns Cul4a Cwc15 Cxxc1 Cyb5 Cyb5b Cyc1 D030016E14Rik D430042O09Rik D5Ertd579e D5Wsu178e Dap Dazap2 Dci Ddx24 Def8 Dgka Dhcr7 Dhrs1 Dhx40 Dis3l2 Dmwd Dnaja2 Dnajb12 Dnajc14 Dnajc25 Dnm2 Dolpp1 Dom3z Dpm1 Dscr3 Dsg2  30 43 10 4 44 134 8 151 23 340 40 10 7 57 133 112 87 80 21 2 58 5 31 23 16 161 5 8 19 4 17 27 69 12 54  35 17 8 16 34 138 13 332 4 721 58 18 6 51 86 60 68 67 13 11 43 6 62 70 31 113 12 16 47 11 15 41 49 12 42  31 17 3 6 44 116 6 451 15 971 48 14 2 48 69 147 89 70 24 2 57 10 47 51 21 169 13 8 17 15 10 42 63 8 37  17 16 0 9 26 80 5 339 15 226 26 10 7 26 63 84 59 31 10 1 66 4 21 32 8 65 3 5 13 4 7 18 60 8 69  28 52 11 7 27 95 6 575 31 716 31 11 12 30 215 132 85 61 16 6 76 15 39 35 30 108 17 17 19 10 25 23 43 7 68  Dtnbp1 Dtx4 Dus1l Dym Ears2 Ecsit Edem1 Edf1 Eif1 Eif1ad Elmod2 Elovl2 Eml3 ENSMUSG00000059333 Epb4.1l2 Epb4.1l5 Ergic1 Etfa Exoc1 Fabp2 Fads1 Faf1 Faf2 Fam125a Fam149b Fam63a Fam82a2 Fam96a Fbxo42 Fcna Fdx1 Fig4 Fip1l1 Flnb Fmc1  4 8 130 8 5 29 34 173 263 145 17 203 13 18 13 14 195 124 12 6 43 20 84 68 9 8 13 264 35 2 60 3 29 20 48  10 8 176 20 11 130 30 507 277 120 33 254 8 21 30 24 285 105 15 48 36 22 122 78 7 24 21 474 47 116 101 18 39 16 79  5 11 122 15 4 93 38 547 290 83 23 167 6 7 26 18 171 47 10 67 81 12 85 85 8 31 14 497 36 60 71 14 13 18 43  6 7 60 2 3 16 28 62 157 32 30 279 8 4 17 16 90 151 10 12 55 6 44 42 9 15 7 210 20 81 63 5 15 10 18  6 6 228 18 9 69 45 436 196 118 36 297 16 18 53 22 133 206 11 90 78 10 55 72 15 45 21 441 18 105 150 10 22 13 44  194  Fnbp1 Foxa2 Foxo1 Ftsj2 Gabarap Galk2 Galnt2 Gcs1 Gdi2 Gfer Glod4 Gm561 Gne Gng10 Gorasp2 Gosr2 Gpd1 Gpt2 Gramd3 Grcc10 Grina Gsta4 Gstm3 Gstt1 H2-Ke6 Harbi1 Hcfc1r1 Heatr6 Hexb Hexim1 Hibch Hif1an Hnf4a Hsd17b2 Hsdl2  17 37 12 5 79 24 22 45 200 14 35 32 10 22 80 8 5 56 15 118 26 13 16 75 29 14 12 13 11 35 12 25 85 14 34  47 60 8 12 71 17 23 73 104 38 31 52 30 13 186 9 26 57 13 82 126 8 44 234 53 7 4 13 88 25 29 33 113 61 30  35 64 8 4 51 7 9 46 224 44 28 21 42 21 90 17 36 66 15 99 160 11 87 404 91 15 8 9 61 29 26 24 91 167 53  33 13 9 1 99 25 8 13 133 5 13 14 10 8 41 4 28 77 20 77 118 22 32 174 44 14 14 11 40 16 15 10 61 103 55  33 102 11 13 92 34 10 46 350 28 28 23 32 15 72 16 46 86 28 74 203 22 75 356 83 25 17 8 44 36 29 20 103 174 71  Hyal2 Ibtk Icmt Ide Ifi47 Ifrg15 Ift74 Ilvbl Imp3 Impa1 Inpp5f Irf6 Irf9 Itgb5 Itih1 Itm2b Ivd Jagn1 Jtb Kbtbd4 Khk Kif1b Klhl22 Kpna4 Kpna6 Krt8 Lap3 Lasp1 Lass2 Ldlrap1 Leng8 Letm1 Lig4 Lipc Lman1  46 48 6 100 0 37 14 11 43 146 10 5 7 28 27 1475 19 118 67 25 28 39 10 18 26 134 87 187 62 14 53 67 11 50 22  78 59 13 84 6 21 16 15 24 256 11 14 45 66 210 3384 17 417 80 32 22 46 12 25 26 182 147 227 87 0 164 60 17 319 34  80 76 5 98 14 47 7 7 19 222 10 7 21 66 203 3239 27 286 61 17 47 47 10 15 17 116 74 229 61 20 98 54 7 303 29  15 43 3 30 4 22 5 9 16 102 13 7 10 25 182 2288 25 40 45 9 37 25 6 13 13 89 34 147 40 11 121 19 2 183 22  42 38 13 111 13 32 7 22 67 281 14 14 34 65 270 3784 41 248 103 12 47 24 11 15 18 172 62 122 107 15 121 59 7 306 18  195  Lman2 Lmbrd1 LOC100044376 LOC100047249 LOC100047674 LOC100047920 LOC631966 Lonp1 Lonp2 Lrfn3 Lrrc3 Lrrfip2 Lyrm2 Mad2l1bp Mad2l2 Manba Map2k4 Map3k4 Mapre3 Masp2 Mcart1 Mcts2 Med4 Mllt4 Mlx Mon1a Mrpl33 Mrpl52 Mrpl53 Mrps31 Mrps35 Mtch1 Mterfd2 Mtf1 Mtm1  83 11 44 28 22 81 6 220 33 12 13 20 13 22 8 10 47 10 9 14 18 37 23 15 31 4 123 272 37 28 76 157 14 16 8  80 19 64 47 40 157 11 663 26 9 30 20 24 120 8 10 70 12 19 134 19 65 31 24 23 12 170 784 27 46 70 114 22 17 45  66 16 50 94 20 118 8 485 10 11 36 12 25 60 6 6 59 19 9 315 10 50 15 14 33 22 157 492 29 22 58 152 7 14 40  33 20 50 30 28 67 5 223 39 4 12 9 9 13 6 10 37 6 5 127 21 9 5 5 19 4 46 182 11 15 27 68 4 11 28  116 31 67 85 21 81 10 255 51 14 40 15 17 58 10 11 54 13 21 318 18 24 19 13 39 21 78 251 27 52 55 125 28 18 62  Mtmr1 Mtmr14 N4bp1 Nagk Napg Nat6 Nbn Ndfip2 Ndst2 Ndufa13 Ndufa2 Ndufa6 Ndufb10 Necap1 Nicn1 Nipa2 Nkiras1 Nosip Npdc1 Npepl1 Nploc4 Nqo2 Nr1h2 Nr1h3 Nr2f6 Nt5dc1 Nudcd3 Nudt16 Nudt8 Oaf Oma1 Oraov1 OTTMUSG00000000757 Oxnad1 Papss1  16 29 15 11 14 5 11 33 14 59 138 591 154 73 11 11 8 6 4 10 42 35 24 5 334 7 25 11 14 28 43 6 10 9 51  48 32 19 74 16 33 12 35 22 33 117 1079 163 150 8 18 9 7 12 26 53 73 60 13 617 42 8 7 42 61 76 10 27 22 55  25 27 25 101 10 53 8 23 21 52 111 1497 124 74 12 7 12 20 5 21 49 165 40 15 666 29 17 11 60 33 79 7 30 12 23  22 20 9 34 17 6 4 39 11 21 51 545 55 49 5 12 10 5 3 7 19 49 32 6 226 8 13 15 4 24 36 4 4 6 16  16 26 15 83 11 74 6 28 41 70 80 1178 300 93 9 10 10 15 4 20 25 137 25 25 377 24 36 17 42 76 44 5 37 31 55  196  Park7 Parn Pccb Pcnxl3 Pcp4l1 Pdcd10 Pdcl Pdcl3 Pdha1 Pdhx Pdxdc1 Pdzk1 Pef1 Pex3 Pex6 Phf6 Pi4k2a Pi4ka Pip5k3 Plekhb2 Plxna2 Plxnb1 Pmpcb Poldip2 Polg Poll Pols Pon2 Por Ppil4 Ppp3r1 Prdx1 Prdx2 Preb Prkab1  303 17 17 13 1 12 50 73 107 17 33 7 10 7 28 11 9 44 4 10 6 147 15 129 8 10 29 23 77 10 59 71 325 50 25  253 32 27 19 8 21 30 105 133 9 44 49 17 1 38 21 9 76 9 20 14 566 30 269 15 17 37 66 105 23 173 207 260 48 40  208 23 20 9 6 16 22 63 186 21 75 46 16 6 30 7 7 37 5 6 9 406 23 262 5 17 57 61 186 7 148 158 278 65 41  85 12 25 6 2 14 42 70 151 12 19 37 10 8 21 6 5 37 11 14 11 290 12 85 6 4 24 31 133 4 29 44 100 36 26  321 12 25 11 12 8 63 41 195 12 92 58 22 12 16 10 11 28 13 12 14 282 14 155 6 7 64 104 245 16 118 171 333 74 23  Proz Prr13 Psmb10 Psmc2 Psmc5 Psmd2 Ptp4a1 Ptplad1 Ptpn12 Pts Pttg1ip Pxmp3 Pygo2 Qprt R3hdm2 Rab11fip1 Rab18 Rab28 Rab35 Rab8a Rad23b Raf1 Rap2a Rasl11b Rassf7 Rbm39 Rchy1 Reep6 Rfk Rfwd2 Rhbg Rhot1 Rhot2 Ring1 Riok2  8 24 177 163 104 146 42 96 57 25 40 41 59 73 45 21 110 31 43 67 154 29 11 9 28 272 20 63 86 31 23 27 35 13 28  29 15 332 201 97 95 248 39 109 35 243 70 105 90 53 31 162 27 83 76 58 50 15 13 35 304 25 144 85 35 22 45 37 11 38  47 26 162 231 122 110 122 96 91 31 211 59 67 144 45 45 202 23 68 66 56 42 6 7 62 155 28 152 92 23 37 42 31 12 26  34 27 49 55 34 32 68 93 39 21 89 47 33 54 41 13 128 19 25 46 60 25 18 4 15 264 15 85 71 19 4 24 24 9 21  54 36 151 111 91 151 191 94 149 21 133 52 41 174 86 51 231 48 80 51 197 72 16 6 40 216 17 165 59 16 29 36 25 11 30  197  Riok3 Rnasek Rnf103 Rnf11 Rnf115 Rnf128 Rnf145 Rnf167 Rnf181 Rnf20 Rnf4 Rnf43 Rock1 Ropn1l RP23-143A14.5 Rpl7l1 Rpp40 Rps6ka4 Samd8 Sap18 Saps3 Sbf1 Sbno1 Sc4mol Scfd2 Sdhc Sec24c Sec63 Serinc1 Setd2 Setd4 Sft2d2 Sgsm2 Sgta Slc10a3  22 29 39 14 28 138 42 24 9 31 170 9 64 6 15 150 29 2 29 11 46 12 34 80 5 110 26 47 7 7 16 46 5 66 22  24 51 253 9 12 233 39 89 8 65 241 19 51 24 19 147 51 6 40 13 72 34 36 382 9 66 23 42 22 17 11 63 8 81 25  16 79 110 11 12 97 31 51 11 30 249 24 56 39 17 166 43 8 31 7 55 39 20 251 11 51 24 37 10 7 7 90 5 104 34  24 31 73 14 14 238 14 30 14 35 117 10 75 10 5 39 6 6 35 7 39 12 31 258 1 93 18 47 12 11 18 55 2 33 12  44 58 152 15 41 299 21 50 15 23 187 20 63 29 18 107 38 10 56 13 39 29 21 285 6 138 42 72 15 8 24 71 12 88 16  Slc12a7 Slc16a10 Slc16a12 Slc16a7 Slc19a2 Slc22a23 Slc25a10 Slc25a15 Slc25a23 Slc25a3 Slc25a45 Slc30a5 Slc30a9 Slc31a1 Slc31a2 Slc35a5 Slc39a3 Slc46a1 Smad3 Smcr8 Smg5 Smoc1 Snap29 Snupn Snx3 Spcs3 Spna2 Spop Spryd4 Spsb2 Srxn1 Stat1 Stim2 Stip1 Stk38  25 13 11 12 55 36 15 23 33 75 17 48 61 112 5 14 4 13 13 11 23 51 11 7 178 24 21 23 24 1 2 8 7 217 32  83 20 31 17 66 55 23 27 58 155 49 46 61 128 26 22 14 8 23 12 16 61 16 11 247 47 105 25 58 29 13 35 8 141 42  46 16 16 12 43 43 27 25 71 124 18 39 49 166 26 29 19 13 15 10 17 41 19 9 239 19 53 4 49 12 3 23 10 233 70  32 7 14 17 70 34 11 12 67 48 19 39 52 169 7 18 3 9 9 12 7 28 8 6 136 10 34 15 33 5 2 22 5 51 37  27 11 12 27 74 22 16 20 111 74 39 38 38 184 32 28 15 11 18 13 15 80 20 10 157 24 70 22 38 14 4 28 9 161 58  198  Stx8 Suclg1 Supt5h Surf4 Syvn1 Tada1l Tanc1 Tbc1d12 Tbc1d23 Tbcel Tbl1xr1 Tesk1 Tex261 Tfg Tfip11 Timm44 Tm2d2 Tm7sf3 Tmbim4 Tmem135 Tmem18 Tmem30a Tmem50a Tmem66 Tmem87b Tmub2 Tnrc6c Tom1l1 Tor1a Tor1aip2 Tpd52l2 Tpmt Tprgl Tram1 Trappc5  18 209 31 158 9 61 14 8 1 36 20 12 24 80 20 56 28 7 35 22 11 98 82 85 7 10 24 64 18 4 10 11 47 158 59  5 594 57 167 6 173 41 12 14 74 21 12 17 133 20 142 19 6 35 36 5 73 50 77 24 12 20 63 47 10 13 15 60 167 62  23 411 42 161 5 172 9 7 2 55 26 5 28 88 24 52 45 6 39 25 10 94 70 98 13 3 7 44 23 5 7 10 105 147 93  11 187 24 99 10 50 24 10 3 37 25 4 17 68 13 33 18 4 34 26 15 109 71 79 10 10 28 30 11 4 5 8 36 150 24  17 241 27 111 15 142 16 8 7 114 35 16 34 83 13 79 24 12 25 62 16 178 133 117 19 9 38 42 28 10 7 11 97 116 90  Trib3 Trim3 Trim8 Trmt2b Trp53bp2 Tspan31 Tst Ttll5 Tuba4a Tubb2a Tubg1 Tusc2 Txnrd3 Ubap1 Ube2q1 Ubqln1 Ubr2 Ubxn4 Ubxn6 Ufsp2 Ugt2b34 Uqcr Uqcrq Uso1 Usp15 Usp19 Usp38 Usp4 Usp47 Vapb Vps37a Vrk3 Wdr45l Xpa Yipf3  4 3 19 11 8 52 160 8 19 18 11 9 26 22 59 77 46 29 22 113 12 820 220 17 31 6 18 40 7 162 10 14 34 10 16  14 7 25 9 13 103 461 12 32 32 11 7 34 29 34 164 63 46 41 187 207 2473 405 21 78 6 24 45 8 217 9 48 30 12 17  11 5 15 4 2 68 346 3 36 26 6 8 46 28 51 293 44 47 43 90 219 2252 347 14 30 3 20 38 13 231 10 54 26 14 20  9 4 14 6 3 102 194 5 11 34 2 3 9 13 40 57 62 35 21 74 124 459 128 15 13 5 12 24 7 112 8 11 26 8 10  19 10 19 11 5 99 452 8 20 38 8 12 23 21 43 240 61 44 22 127 275 1254 691 10 39 11 21 49 12 110 7 47 21 19 20  199  Zc3h13 Zc3hav1 Zcwpw1 Zfand2b Zfc3h1 Zfhx4 Zfp187 Zfp259 Zfp319 Zfp330 Zfp358 Zfp444 Zfp637 Zmat2 Zyg11b Zyx  6 10 5 19 14 15 2 23 13 6 12 12 12 74 30 36  11 53 10 91 15 17 37 13 19 12 6 15 16 100 32 198  6 44 5 108 10 10 10 19 18 3 9 9 7 101 23 86  5 14 7 30 17 13 18 11 13 5 5 8 7 33 26 73  5 46 5 120 13 14 18 20 10 4 17 7 11 53 18 87  200  Appendix F Cluster F genes and their expression in the five Tag-seq libraries (in tags per million).  Symbol 0610031J06Rik 1300001I01Rik 1700021K19Rik 2410001C21Rik 2610008E11Rik 3110003A17Rik 4933426M11Rik 6330416G13Rik 9030625A04Rik A230067G21Rik a2ld1 AA987161 Aadac Aass Abhd5 Acaa2 Acadl Acadvl Acly Acp2 Adck4 Adh5 Adssl1 Afp Aga Agl  E10 DLK1+ 73 113 15 17 8 36 8 6 20 9 22 5 5 82 18 362 107 22 250 15 5 165 36 8853 33 9  E12 DLK1+ 315 228 31 15 10 107 13 6 48 27 23 11 28 118 15 799 346 13 796 30 4 165 51 14795 29 25  E14 DLK1+ 305 266 31 18 6 125 14 8 40 19 29 12 92 99 24 1412 281 67 619 37 14 164 129 29857 39 12  E16 DLK1+ 107 94 25 22 13 57 11 11 49 19 54 10 51 156 32 1048 295 61 571 27 15 223 30 24624 63 18  adult liver 147 111 10 18 10 50 10 4 13 13 52 10 36 91 13 767 184 41 245 14 2 158 86 0 29 13  Aif1 Akap9 Alpl Anxa2 Ap3d1 Apoa1 Apoc2 Araf Arfgap1 Arhgap18 Arhgap9 Arhgef11 Arhgef3 Arntl Arsa Atg9a Atox1 Atp6v1b2 Atp6v1c1 Atp6v1e1 Atp9a B230312A22Rik B4galt1 Baz2b BC004004 BC021614 BC024659 Bet1l Blvrb Bnip2 C1qb C1qc  1 10 6 32 40 1244 81 34 5 3 3 5 3 5 6 8 185 35 42 70 7 5 66 13 36 2 9 3 22 58 2 0  42 13 15 89 47 5257 211 65 6 16 17 11 9 4 16 2 274 58 39 68 6 5 90 18 22 19 19 3 122 77 113 12  43 12 15 219 78 4871 524 91 13 24 12 5 10 16 16 12 415 84 84 133 24 5 116 8 63 177 21 13 157 155 60 2  76 13 9 216 42 5122 474 45 11 23 16 12 13 17 12 8 245 40 40 95 12 11 72 30 53 140 11 10 81 146 116 12  23 8 4 3 39 2858 57 26 2 4 0 6 15 19 5 6 293 34 38 63 19 10 49 21 25 4 13 13 98 52 72 7  201  C2 Cab39 Cd302 Cd37 Cd47 Cd86 Cdh1 Cfp Cldn1 Cln8 Cmklr1 Coro7 Cpb2 Cpox Cpt2 Crat Csgalnact2 Ctla2b Ctsd Ctsl Cybrd1 Cyhr1 Cyp51 D15Ertd621e D430028G21Rik D6Wsu163e Dbt Ddrgk1 Ddx28 Deb1 Dennd3 Derl2  8 66 42 0 69 0 31 1 14 18 1 7 49 230 44 15 11 0 65 259 3 21 87 168 30 7 136 59 18 96 2 58  19 79 83 15 128 19 141 20 65 15 5 8 99 245 60 46 16 6 181 357 11 54 91 196 42 12 166 83 36 122 4 101  22 63 155 16 285 19 129 23 66 29 7 6 144 276 62 73 13 11 229 479 12 65 115 240 35 14 113 225 50 141 11 133  42 104 143 11 211 30 94 35 113 29 11 12 185 431 67 28 19 8 219 383 10 21 129 244 49 8 245 55 13 114 6 122  14 46 65 7 83 5 55 22 27 10 4 9 74 444 30 31 17 7 64 274 0 29 83 224 41 6 168 119 19 66 5 108  Dhcr24 Dnase1l1 Dock8 Efha1 EG433273 Eif2b2 Enpp4 Enpp5 Esrra Etfdh F10 Faah Fam120b Fam129b Fbxo8 Fcgrt Fech Fem1a Fgg Fggy Fhod1 Flad1 Flot1 Fndc3a Folr2 Frmd4b Ftcd Galm Gcc2 Gcgr Gchfr Ggt1  270 2 6 5 2 26 20 30 5 48 106 12 19 11 7 29 22 14 361 7 5 5 7 62 0 6 13 12 27 9 27 4  580 9 9 9 8 35 12 56 10 52 642 31 19 44 8 52 82 65 1548 10 15 6 31 237 9 8 62 16 95 69 63 4  903 24 18 11 18 63 38 56 12 87 725 32 41 53 9 79 81 88 1925 20 28 12 35 174 12 6 79 25 102 138 79 21  568 16 18 10 0 10 24 53 5 79 209 47 23 22 16 90 56 8 1932 24 21 7 14 162 12 16 77 27 34 167 68 18  601 9 13 3 8 27 29 21 5 85 377 52 11 23 12 94 56 38 2389 28 17 6 21 101 6 12 61 13 49 63 43 0  202  Gjb1 Glce Glrx Gm2a Gng11 Gng12 Golga4 Golt1a Gpam Grb14 Grhpr Gstz1 Gys2 Hadha Hdac10 Hdhd3 Hectd3 Hgd Hgs Hint3 Hist2h2be Hs1bp3 Hsd17b11 Htatip2 Ifi35 Igdcc4 Ikbke Ikbkg Insr Irf3 Itfg3 Itga1  7 10 37 20 15 27 34 15 4 47 176 153 16 25 8 11 6 30 5 5 11 4 32 6 0 7 14 5 14 19 29 3  16 11 135 38 33 75 111 58 12 49 372 565 53 17 19 26 10 224 4 6 19 5 79 3 0 17 64 10 54 35 50 18  30 15 147 47 51 89 103 103 13 71 496 982 57 55 18 36 18 280 11 12 25 11 63 14 11 30 98 18 33 51 52 20  23 13 91 51 59 72 75 13 7 83 151 382 51 54 18 11 11 192 5 11 24 7 87 11 12 24 98 12 73 30 38 25  28 7 53 52 4 10 89 38 4 62 266 340 42 29 4 15 16 244 4 7 8 10 33 7 4 0 10 8 66 37 37 17  Itpr3 Iyd Kank3 Kif16b Klhl6 Lbp Lcorl Lims1 LOC100044703 LOC100044830 LOC100047837 Lrig2 Lrrk1 Ltbr Lyl1 Lypla1 Man2b2 Maob Marco Mbc2 Mbl1 Met Mfsd1 Mfsd7c Mgmt Mif4gd Mosc2 Mpnd Mst1 Mthfd1 Myl9 Nab1  7 5 4 5 3 21 6 6 16 111 9 5 6 85 4 93 4 88 0 14 5 57 32 2 8 5 68 7 4 244 9 21  8 14 17 11 7 92 9 15 20 100 17 1 12 267 10 204 12 80 107 38 37 64 26 8 11 9 76 4 45 237 52 51  19 38 12 9 14 341 11 7 20 180 24 12 30 525 27 212 12 182 80 60 37 126 54 13 15 24 165 17 51 328 74 63  19 30 19 7 8 168 7 14 19 121 28 12 15 73 13 107 17 169 85 33 49 85 51 5 13 12 91 7 77 250 62 82  0 43 4 3 5 165 5 4 12 70 14 0 5 225 0 102 4 44 7 23 47 111 56 10 7 7 108 11 22 197 7 62  203  Naga Nags Nckap1l Ndel1 Nf1 Ngly1 Ninj1 Nme6 Npr1 Nptn Nrbp1 Nsdhl Nudt13 Nudt18 Nxt2 Ofd1 P4hb Pafah2 Pak1 Parva Pcbd2 Pcdh1 Pcdh24 Pcsk4 Pcyox1 Pde6d Pex11c Pex16 Pf4 Phf3 Phlda2 Phyhd1  15 8 0 46 9 8 19 138 4 53 38 17 6 2 7 9 258 16 6 20 73 10 13 18 79 7 7 10 4 12 14 26  27 4 15 95 7 7 11 333 11 54 84 31 10 4 10 19 460 21 11 25 26 12 24 61 181 7 13 12 18 13 48 68  27 14 10 119 11 18 41 423 34 61 138 21 12 25 17 20 684 43 28 29 92 9 30 44 323 17 23 54 36 11 39 47  25 15 15 97 13 15 27 130 13 69 23 39 19 8 16 16 374 39 23 51 86 15 24 65 149 4 14 29 35 17 37 70  9 12 3 58 6 14 12 160 24 35 53 20 11 5 6 9 321 30 0 54 49 7 0 36 163 8 10 38 0 10 0 57  Pkp2 Plac8 Pld1 Plekho2 Plg Plod1 Pmvk Pnpo Pold4 Ppox Prkca Prodh2 Psd4 Ptgr2 Ptrf Pxmp4 Rab17 Rab3d Rabac1 Rabgap1l Rad51l1 Rag1ap1 Rcl1 Reep3 Rela Rhbdf1 Rhog Ripk1 Rnh1 RP23-100C5.8 Sardh Scamp1  43 8 3 2 11 13 22 25 4 18 7 43 0 33 1 21 6 4 20 3 6 17 95 7 8 6 9 5 25 7 20 7  209 19 21 19 27 21 63 19 7 71 14 819 2 33 11 29 5 14 13 6 8 44 87 8 9 14 9 11 35 6 55 8  266 23 9 16 51 21 111 40 14 68 38 1859 11 49 26 29 18 22 27 7 14 36 139 13 22 20 12 8 65 6 213 15  164 26 18 13 44 18 56 37 9 51 24 401 3 43 26 54 11 11 25 11 10 27 152 18 10 15 19 14 24 14 53 8  189 23 4 4 11 9 18 13 6 15 12 890 3 36 8 50 3 6 13 11 2 15 73 15 10 5 9 9 27 14 103 10  204  Sdpr Sec11c Sec24a Sema4g Sepw1 Serpina10 Serpinc1 Serpind1 Sertad1 Slc13a3 Slc25a12 Slc25a39 Slc25a46 Slc27a4 Slc2a2 Slc33a1 Slc35a3 Slc35b4 Slc35f5 Slc38a7 Slc39a4 Slc44a3 Slc6a13 Slc6a19 Slc7a8 Slmap Smap2 Smox Snx14 Sord Spon2 Spp2  20 23 85 70 236 35 581 61 8 9 8 380 26 7 19 51 5 7 52 8 3 2 10 1 12 16 15 6 35 349 8 155  125 49 285 116 329 42 1992 130 11 67 15 285 19 7 86 78 4 11 52 6 7 11 42 5 29 44 11 12 53 359 54 703  166 67 199 256 651 62 2551 171 15 76 20 733 25 27 73 70 8 21 55 12 18 49 86 10 24 17 28 14 51 576 80 961  120 31 204 300 112 64 2408 93 12 71 9 293 31 4 72 77 12 8 88 8 8 37 35 8 24 37 28 15 56 821 42 1069  21 33 196 140 299 28 1054 96 7 14 8 256 19 9 16 36 5 13 42 6 10 2 55 0 21 13 27 2 35 445 18 158  Srebf1 Srgn Stambp Stard10 Stat6 Stbd1 Stx2 Stx5a Stxbp2 Stxbp3a Svil Tbc1d15 Tbc1d17 Tbc1d22a Tbk1 Tcta Tdrd7 Tex2 Tfpi2 Tgfb1i1 Tgfbr2 Tgm2 Timd2 Tmem150 Tmem63a Tmem86b Tnip2 Tpk1 Tpst2 Trafd1 Trap1 Trf  102 2 9 709 1 3 6 53 5 36 14 44 6 24 36 8 3 17 19 4 20 69 3 53 3 34 3 7 15 8 20 2700  129 15 7 2467 17 42 13 77 13 50 11 48 13 24 69 10 21 20 39 27 55 114 19 105 19 31 6 11 17 13 30 17574  178 15 14 2500 12 47 9 151 14 63 30 52 14 45 34 19 21 35 83 27 37 185 24 135 10 78 12 3 24 21 31 23582  206 19 19 917 15 53 10 122 9 63 34 89 7 42 77 14 18 40 66 29 74 99 25 62 21 61 9 17 18 21 23 16670  214 17 10 1154 8 48 4 27 8 22 15 77 5 6 37 0 20 41 17 3 22 53 12 65 7 48 8 6 18 26 16 5544  205  Tsga14 Tspan7 Tspo Txndc10 Ubtd1 Ucp2 Vamp8 Vat1 Vkorc1 Vps41 Wipf1 Yif1a Zc3h7a Zcchc24 Zfp524 Zfp704 Zfyve26  12 238 20 19 2 9 48 17 62 7 24 13 14 19 15 16 25  22 813 31 25 6 6 80 23 71 10 44 43 62 14 25 13 44  24 865 71 14 12 23 197 35 156 18 41 112 23 27 31 20 124  12 315 55 38 2 17 67 34 108 12 45 15 58 40 19 29 52  12 427 73 35 8 2 93 9 54 10 17 52 13 28 12 16 28  206  Appendix G Cluster G genes and their expression in the five Tag-seq libraries (in tags per million). Symbol 1200009I06Rik 1810043H04Rik 2510049J12Rik 5133400G04Rik 5930434B04Rik Abcc4 Ache Acot8 Acss1 Aftph Akr1c13 Aldh1b1 Ano10 Ap1b1 Ap2m1 Ap4s1 Arl5c Athl1 Atp5a1 Atp5b Atp5j2 Bag1 Bbs5 Bcl2l13 Brf2 Ccdc71 Cenpt Clec1b Clstn3 Cog4  E10 DLK1+ 18 8 7 4 8 2 0 2 26 13 76 9 6 77 18 52 8 12 1741 2100 256 25 9 76 7 16 6 2 7 3  E12 DLK1+ 68 34 18 15 5 6 9 4 72 35 205 88 5 136 21 99 46 59 4210 6626 452 57 18 171 9 27 16 252 10 15  E14 DLK1+ 52 20 49 14 24 13 27 14 162 30 202 85 29 165 63 93 60 59 3295 5813 426 44 15 139 13 41 20 223 22 14  E16 DLK1+ 25 5 12 4 7 5 3 2 82 20 121 35 11 61 24 38 28 23 1182 1663 144 16 5 58 2 6 7 51 4 3  adult liver 0 7 11 0 4 0 0 2 0 8 49 8 0 22 4 23 0 22 1085 828 117 9 5 53 2 7 0 35 3 3  Copz1 Coq3 Ctla2a Ctps2 Ctr9 Ctsz Cyba Dgcr8 Dpagt1 Dtnb Egln2 Ehd4 Eri3 Ero1l F12 F2rl2 F2rl3 Fam109a Fastkd5 Fbxo36 Fermt3 Fkbp11 Gale Galk1 Galnt4 Gata1 Gbf1 Gfi1b Gmppb Gp5 Gp9 Gsn Gstt2 Hebp2 Hip1r  114 11 1 116 5 125 6 21 35 6 61 9 10 10 161 2 0 18 12 7 2 95 16 410 26 1 45 1 6 0 0 53 8 18 20  260 23 33 343 5 521 29 34 61 17 311 33 16 29 754 105 7 103 15 56 11 458 40 515 79 8 136 15 25 13 12 180 105 26 61  292 23 19 229 11 632 30 43 114 16 352 43 40 27 930 154 14 109 24 31 21 545 82 1001 69 10 101 24 26 24 31 159 150 31 61  97 7 8 72 3 225 9 12 51 5 63 9 10 16 403 15 0 11 5 3 10 94 15 247 51 4 37 4 7 2 0 124 28 4 26  49 6 0 10 0 120 2 2 14 5 91 6 6 3 160 0 0 17 4 14 0 10 24 20 11 0 21 0 5 0 0 4 24 2 2  207  Hsdl1 Ikbkap Ilk Itga2b Krt20 Lcp2 Lgals3bp Lgtn Lin28b LOC100044423 LOC676596 Lpcat3 Lrpap1 Lrwd1 Mcat Mecr Mogat2 Mrpl37 Mrps11 Mto1 Mvd Myg1 Nckipsd Ncln Ndufs2 Nfkbib Nipsnap1 Nisch Nomo1 Nostrin Npb Nrgn Nsd1 Nt5m Nucb1  10 39 50 0 51 0 47 27 29 6 101 21 31 44 27 33 128 38 34 30 45 55 6 36 498 10 149 104 99 31 12 1 57 9 21  18 93 91 170 299 35 284 76 85 22 118 67 94 61 76 53 229 74 34 72 269 116 7 128 1578 21 367 409 343 68 131 34 174 29 73  28 108 278 223 170 81 444 73 77 14 226 121 77 79 140 68 294 147 83 86 238 187 19 242 1369 30 703 330 230 124 184 35 151 35 54  7 16 72 9 62 19 87 15 7 7 82 34 60 15 15 20 76 25 13 29 91 38 6 27 360 9 120 121 47 41 34 6 59 8 16  9 8 19 3 0 12 52 21 0 5 17 10 7 6 30 11 0 23 3 31 38 17 4 46 461 5 142 62 35 27 0 0 37 6 17  Nudt16l1 Nup43 Ociad1 Ostc P2rx1 Pet112l Pfdn1 Pigyl Pla2g12b Pla2g4a Pld3 Plek Polr1a Ppap2c Ppm1f Psat1 Psmd4 Rabif Rap1gap Rasgrp2 Reep2 Rgs10 Rgs18 Ripk4 Rpusd1 Rsu1 Rtel1 Ruvbl1 Samd14 Scamp4 Sdf2 Senp5 Serpina6 Serpinb10 Serpinb2  122 10 49 204 1 25 88 29 153 3 8 1 9 56 39 976 27 17 9 0 1 0 0 19 11 19 32 318 6 34 118 30 575 0 0  271 30 74 717 29 37 162 55 247 11 19 36 14 81 72 4065 39 41 36 10 25 14 32 66 17 43 68 747 10 72 262 55 1746 2 11  235 20 161 821 82 53 323 77 406 9 26 35 33 127 89 3649 70 31 40 15 19 20 29 67 29 65 75 842 33 86 283 66 2066 16 37  96 3 37 213 5 20 39 16 135 5 8 12 4 45 23 713 15 12 25 4 0 5 6 18 9 21 28 117 5 14 135 34 837 2 4  11 2 27 63 0 13 21 10 73 0 2 0 0 13 19 0 14 9 3 0 0 2 0 7 4 15 28 67 0 23 90 14 113 0 0  208  Sgsh Sigirr Slc30a1 Slc35b2 Slc35c2 Slc35d3 Slc44a1 Smug1 Smyd3 Snw1 Spns1 Sqle Sumo3 Supt6h Syngr1 Sytl4 Tbxa2r Timm22 Timp3 Tmed1 Tmem180 Tmem208 Tnnt1 Tomm70a Tpi1 Ubl5 Vcam1 Vps33b Vwf Wars2 Xk Xpnpep1 Zc3hc1 Zfp13 Zfp768  5 11 35 62 37 1 10 5 11 218 22 78 10 12 9 1 11 43 2 57 13 88 6 246 10 380 41 12 0 26 4 43 76 8 10  11 12 90 114 41 13 37 15 20 371 66 233 14 35 12 12 20 59 20 200 15 134 25 1105 20 759 167 50 21 87 15 117 129 23 20  10 40 112 119 118 15 68 14 18 351 83 309 20 30 22 20 40 114 24 238 33 215 20 659 31 1029 186 34 37 111 12 102 143 20 18  3 16 58 28 47 5 28 4 9 89 11 111 8 10 5 3 13 21 6 22 13 54 6 150 4 286 91 14 8 19 5 26 40 2 2  2 2 27 18 3 0 10 4 0 73 10 61 0 6 0 2 4 8 4 28 0 37 0 217 7 238 7 14 0 23 5 12 21 0 5  Zfyve19  18  43  38  10  14  209  Appendix H Cluster H genes and their expression in the five Tag-seq libraries (in tags per million).  Symbol 0610009O20Rik 0610010E21Rik 0610010K14Rik 0610040J01Rik 1110031I02Rik 1110032A13Rik 1110058L19Rik 1300018I05Rik 1810009O10Rik 2010209O12Rik 2310003C23Rik 2310003F16Rik 2310011J03Rik 2310030N02Rik 2310047O13Rik 2310061I04Rik 2610030H06Rik 2610034B18Rik 2610036D13Rik 2810428I15Rik 2810432L12Rik 2810433K01Rik 2900010M23Rik 3110001D03Rik 4733401H18Rik 4933433P14Rik 5430437P03Rik 8430410A17Rik Aamp  E10 DLK1+ 31 50 61 50 19 18 46 10 30 9 40 637 6 22 34 56 14 11 11 79 20 26 163 105 12 15 84 33 131  E12 DLK1+ 14 42 47 27 19 22 20 3 41 14 28 410 6 14 29 38 23 9 6 72 27 17 209 142 10 15 47 41 105  E14 DLK1+ 50 33 85 48 31 23 54 11 41 14 26 410 10 25 21 46 14 14 10 102 30 29 194 160 12 21 92 35 108  E16 DLK1+ 20 16 33 40 10 10 32 4 8 8 19 278 6 15 21 26 16 12 6 62 11 13 44 50 10 18 44 10 88  adult liver 18 7 9 12 7 7 15 2 12 5 11 111 3 4 12 6 7 4 2 10 0 0 49 54 0 6 21 10 22  Abhd13 Actr1b Actr3 Acy1 Adck5 Afg3l1 Agpat5 Agtrap Ahctf1 AI597468 Aif1l Akr1c19 Alg3 Alg5 Alg6 Alkbh6 Amot Ampd2 Anapc1 Anapc2 Anapc7 Ankrd27 Ankrd54 Anln Anxa4 Ap1m1 Aph1a Apom Apool Aqp11 Arf6 Arfrp1 Arhgdia Arhgef16 Arid3b  19 37 376 25 7 35 54 18 58 40 40 67 66 26 23 54 22 82 67 65 17 50 56 10 59 99 30 470 11 21 302 36 32 17 10  20 43 507 15 4 23 52 5 52 39 31 27 45 39 11 34 16 132 107 77 11 40 13 8 59 111 11 669 8 13 181 19 10 16 4  19 48 413 31 10 29 44 18 56 28 33 46 64 28 32 48 22 109 114 129 14 41 35 12 60 196 32 925 11 18 235 24 34 29 13  9 14 431 21 8 17 20 14 42 39 7 44 19 21 16 31 19 71 49 60 8 31 34 5 55 101 21 517 7 18 222 26 28 19 3  6 4 79 13 4 10 9 0 22 20 0 11 4 13 6 7 4 34 35 27 0 17 8 0 9 11 2 88 4 8 21 10 13 0 0  210  Ascc2 Asf1b Atp13a2 Atxn10 AU014645 Aurkb B3gat3 Baz1b Bbc3 Bcap29 Bche Bcl10 Bex1 Bex2 Bloc1s1 Brcc3 Bxdc2 C1galt1 C85492 Camsap1l1 Cand1 Capn7 Capns1 Cars Cars2 Casc3 Cc2d1b Ccdc123 Ccdc127 Ccdc32 Ccdc5 Ccdc55 Cd2bp2 Cdc2a Cdc37  16 62 18 260 43 41 24 90 78 17 47 27 1506 1060 18 12 67 10 13 22 46 19 48 27 21 39 34 14 40 28 19 11 26 508 54  20 36 14 248 36 38 28 72 133 10 54 17 1607 897 21 10 33 6 10 23 29 9 38 31 17 25 54 15 30 24 13 10 21 415 34  20 53 19 312 34 39 28 61 109 12 66 28 1544 1154 32 4 53 10 12 23 29 21 69 44 15 37 53 16 41 36 22 9 13 369 59  16 27 15 120 13 19 19 42 97 10 67 17 726 689 23 8 27 9 4 17 15 18 49 25 11 34 43 8 27 23 8 11 10 233 27  5 0 6 88 4 0 4 20 46 2 21 10 0 0 4 3 10 4 2 6 10 10 8 16 4 9 16 4 14 8 0 5 4 3 18  Cdca3 Cdca7l Cdk5rap3 Cdt1 Cep57 Cetn2 Cetn3 Cgrrf1 Chchd7 Chchd8 Cherp Chml Chmp6 Clic4 Clptm1l Cmtm4 Col27a1 Commd1 Copb1 Cops6 Coq4 Coq9 Cotl1 Csnk1a1 Ctps Cuedc2 Cwf19l1 D10Wsu52e D17Wsu104e Dap3 Dbnl Dctn2 Dctn6 Ddx1 Ddx39  37 10 63 69 24 36 192 7 21 11 100 5 21 17 40 13 29 177 77 196 10 115 20 168 122 37 16 111 69 98 39 18 74 111 128  12 4 26 68 24 32 100 3 20 12 89 4 19 6 30 7 32 167 69 173 9 185 17 165 105 33 8 125 23 135 25 4 47 120 66  45 9 57 77 21 33 147 12 25 12 88 12 32 18 30 21 27 145 97 239 13 149 22 120 111 25 14 180 65 162 55 15 71 116 110  12 3 49 39 24 15 109 5 15 3 60 6 10 15 14 5 33 111 71 137 2 127 18 113 26 16 10 75 40 54 20 11 58 76 39  0 0 23 2 0 13 30 4 11 6 37 0 8 0 16 6 12 19 37 46 3 38 3 33 2 5 2 19 15 49 7 3 11 10 3  211  Ddx52 Denr Dhx33 Dhx8 Diablo Dnajc11 Dnajc19 Dpf3 Dpm2 Dpp3 Dpy19l1 Dpy19l3 E430028B21Rik Edc3 Ehd1 Ei24 Eif1a Ensa Entpd6 Ephb4 Epn1 Ercc2 Esco2 Etf1 Fam108a Fam40a Fam96b Fastkd3 Fbxo21 Fbxw11 Flii Frap1 Frmd8 Fth1 Fto  14 268 24 13 17 100 40 7 43 44 33 13 31 40 44 69 72 78 39 14 17 43 11 228 127 25 54 42 99 42 26 17 24 872 63  16 222 27 12 15 133 16 8 26 21 17 10 27 54 71 61 79 95 28 18 22 51 8 203 204 17 45 32 85 23 17 13 30 413 38  18 230 23 11 10 137 44 10 53 42 20 11 28 50 75 84 44 70 39 16 15 59 8 177 194 20 39 37 53 37 35 14 23 715 59  8 132 11 8 10 48 38 4 28 25 21 7 13 35 41 40 57 86 16 6 13 27 7 89 84 17 21 27 22 30 28 12 14 725 46  2 48 10 2 5 41 7 0 13 10 5 0 14 3 9 34 32 27 3 6 7 8 3 91 49 4 9 18 35 16 12 0 12 293 10  Fyttd1 G6pdx Gaa Ganab Gcsh Gdap2 Gga3 Glb1 Glt8d1 Gltpd1 Gm1040 Gmpr2 Gna11 Gnb2 Gnb2l1 Gpatch1 Gpn2 Gpr89 Gpsn2 Gpx7 Grik5 Gtf2f1 Gtf2h5 Gtf3c6 Gtrgeo22 H13 Habp2 Hace1 Herc4 Hice1 Higd2a Hlf Hmg20b Homez Hpcal1  52 45 22 66 402 20 29 25 23 9 20 21 55 136 1900 19 32 36 250 7 60 31 45 116 30 69 228 16 20 10 127 32 31 15 25  42 31 17 12 137 12 48 22 18 8 15 18 32 170 992 24 25 42 388 6 83 31 42 97 20 17 323 20 23 2 47 36 22 13 20  37 41 17 75 418 29 43 13 10 10 15 17 58 160 1849 31 18 45 516 14 70 44 83 150 19 104 287 22 18 10 129 29 15 9 19  33 15 19 53 297 26 29 17 15 5 7 18 22 135 689 13 7 24 282 6 48 13 29 48 12 84 195 18 22 6 113 23 20 1 21  16 0 6 14 83 9 9 5 4 0 2 2 10 60 205 12 7 10 76 0 24 8 11 12 5 5 84 5 11 0 29 13 5 3 0  212  Hprt1 Hscb Hspa5 Ihh Ik Incenp Ing2 Ipo13 Itga5 Jarid1c Jmjd7 Josd2 Jup Katna1 Kbtbd2 Kctd17 Kctd20 Kdelc1 Kifc3 Klhdc10 Klhdc2 Klraq1 Kras Krtcap2 Lars2 Lect2 Leo1 Leprotl1 Lgals2 Lig3 Lin7c Lmna Lrpprc Lrrc14 Lsg1  164 11 1110 5 153 67 17 15 34 25 30 11 36 34 66 29 36 36 23 32 88 11 93 52 9 61 93 33 8 24 37 13 134 18 32  217 8 572 9 123 63 9 9 31 18 31 4 42 38 90 16 11 29 23 23 85 5 108 48 2 31 54 23 7 17 35 6 139 14 33  218 9 1107 10 165 54 16 14 27 19 46 8 33 30 77 22 34 23 9 25 111 14 103 66 16 70 70 26 12 18 29 11 203 12 33  89 8 841 7 142 34 10 5 23 21 25 5 18 20 87 16 23 17 18 23 49 11 108 35 5 41 67 18 3 17 31 10 192 4 20  81 4 91 0 57 0 6 4 5 7 12 3 15 12 31 3 5 5 6 14 28 3 9 7 2 26 35 9 0 4 19 3 23 3 7  Maea Magohb Mak10 Mak16 Map3k11 Mapk1 Mapk1ip1l Mat2b Maz Mbd6 Mbtps1 Med1 Med11 Med21 Med9 Mettl6 Mettl9 Mfsd5 Mkln1 Mms19 Mphosph6 Mrpl10 Mrpl15 Mrpl17 Mrpl2 Mrpl28 Mrpl30 Mrpl35 Mrpl36 Mrpl40 Mrpl50 Mrps12 Mrps15 Mrps23 Mrps25  97 11 27 66 35 237 11 14 63 20 171 65 11 121 23 17 292 15 23 17 53 95 38 117 152 123 158 9 42 37 88 206 114 104 64  57 4 29 50 44 241 12 12 20 17 216 87 6 132 22 19 348 14 14 15 48 79 51 101 213 92 157 11 49 27 65 198 102 107 46  67 11 22 45 35 219 7 16 42 24 275 55 15 137 27 20 334 16 15 15 39 143 47 77 252 127 120 7 39 43 34 332 75 128 45  62 5 15 26 21 146 8 12 40 18 122 64 6 84 13 13 184 9 16 17 20 37 12 37 63 49 84 5 16 12 58 110 42 35 23  12 0 5 15 9 57 3 8 11 10 73 31 0 63 3 0 94 4 7 3 4 21 18 25 64 56 53 4 13 5 17 41 38 46 7  213  Mrps9 Mtap Mtap4 Mtmr12 Mtmr4 Mtpn Mtx1 Muted Myst3 Narfl Ndufa10 Ndufb9 Ndufc1 Ndufs4 Ndufv2 Nedd8 Nek6 Nhlrc2 Nle1 Nln Npc2 Nr6a1 Nubp2 Ola1 Orai1 Paip2 Palm Pbk Pcbd1 Pdcd2l Pdia5 Perld1 Pex14 Pfn1 Pgls  59 194 35 18 59 198 11 17 14 10 564 394 175 154 91 257 89 21 19 14 93 38 30 167 12 98 19 56 415 48 69 41 42 338 85  34 160 20 18 48 166 13 15 15 11 380 238 234 144 62 78 62 8 15 12 132 41 17 120 23 78 16 57 285 57 81 30 50 288 42  48 220 60 25 33 186 11 15 9 8 497 516 151 235 67 222 48 19 21 18 179 29 28 122 28 105 15 49 645 43 126 40 72 311 86  39 38 45 8 37 194 10 13 13 5 378 347 137 136 49 119 56 14 7 13 104 37 24 87 17 63 6 23 496 31 96 24 41 123 45  15 56 7 4 9 86 0 4 4 5 301 94 83 37 32 29 19 7 6 7 33 7 5 50 7 37 0 0 111 12 9 4 20 20 7  Pgpep1 Phldb2 Pias3 Piga Pip4k2b Pknox1 Pla2g6 Plcb3 Plekhj1 Pmf1 Pmpca Pnpt1 Pofut2 Pole3 Polr2g Polr2j Polr3d Polrmt Ppa1 Ppie Ppif Ppil3 Ppm1a Ppp1r11 Ppp1r16a Ppp5c Pppde2 Ppt1 Prkaa1 Prkra Prmt7 Prrc1 Psmb1 Psmd6 Psmd7  8 49 9 20 47 23 46 31 169 59 71 32 29 26 168 50 71 15 453 17 109 12 26 23 11 19 22 20 45 66 54 19 334 92 198  6 86 15 24 33 19 27 17 212 34 31 53 37 19 166 64 70 11 167 15 90 10 13 17 2 11 15 16 46 91 56 23 501 72 84  10 76 15 27 58 24 42 32 206 58 66 57 37 23 130 61 61 14 453 13 111 10 38 16 11 14 18 12 35 70 62 22 478 78 174  6 39 10 18 27 19 27 28 69 26 34 25 15 15 88 16 58 10 193 8 41 4 32 13 5 11 7 10 28 41 16 18 139 35 69  5 27 3 5 23 9 11 16 24 7 21 13 6 6 12 10 10 2 114 4 14 3 5 7 3 3 4 6 3 30 6 7 110 26 48  214  Psmd9 Psmg1 Psph Ptp4a3 Qars Qser1 Rab3gap1 Rac1 Rad1 Rala Ranbp3 Rars Rbm26 Rbm38 Rcbtb1 Rell1 Rer1 Rfc1 Rgl2 Rhbdd2 Rhbdd3 Ric8 Rilpl2 Rint1 Rnf34 Rnpep Rpl4 Rplp0 Rpn1 Rpn2 Rps4y2 Rraga Rras2 Rufy1 Rxrb  43 40 81 19 99 48 11 857 36 21 115 132 87 69 56 30 181 47 21 19 18 18 9 6 13 19 14 11 615 379 81 82 72 14 28  44 31 77 18 61 43 6 801 34 9 112 107 68 72 52 30 92 34 12 27 15 13 8 8 10 10 15 9 543 253 38 82 103 11 38  31 29 113 20 104 29 10 872 29 18 80 160 63 71 57 30 219 25 16 30 12 21 11 11 9 17 13 9 609 370 100 92 89 15 34  35 14 47 15 52 27 7 711 9 8 66 62 54 35 50 28 121 35 16 15 15 14 10 5 10 16 10 7 393 293 42 47 58 14 15  6 3 3 6 6 10 2 338 13 2 15 34 11 0 18 14 80 13 5 7 8 10 3 2 0 3 0 3 122 65 11 35 35 4 11  Samm50 Sass6 Scly Sdf4 Sdr39u1 Sec22b Sec61a1 Serp1 Serpinf2 Setdb1 Sf1 Sf3a1 Sf3b5 Sfrs9 Sh3bp5 Sh3bp5l Sh3gl1 Shmt1 Shmt2 Sin3b Skiv2l2 Slc25a13 Slc27a3 Slc29a3 Slc30a6 Slc35a2 Slc35b1 Slc35e1 Slc38a10 Slc39a11 Slc9a1 Slc9a3r1 Snrnp200 Snx24 Snx5  58 13 12 63 7 56 443 436 314 48 14 42 141 132 60 11 56 63 64 82 92 111 13 7 18 8 199 27 21 20 13 215 124 71 229  38 13 13 83 1 38 372 504 369 36 9 26 78 91 14 13 50 77 71 122 105 60 12 14 21 14 171 16 10 18 18 155 71 92 227  73 11 9 104 11 37 372 436 399 41 18 44 126 100 68 14 33 75 44 132 94 121 18 13 28 13 264 17 30 15 20 128 76 56 280  39 7 10 48 7 43 377 457 224 39 15 13 26 53 52 6 23 45 47 72 48 108 4 9 17 9 69 18 22 17 16 159 70 74 164  17 0 5 23 0 12 31 133 102 16 0 7 31 14 19 4 16 17 20 27 34 31 0 0 2 2 116 5 9 3 7 32 29 29 105  215  Snx6 Spcs1 Sptlc1 Srm Ssr4 Ssu72 St13 St8sia3 Stk25 Stoml2 Stx18 Surf1 Suv39h1 Sypl Sys1 Taf6l Tagln2 Tank Tars2 Tbc1d9b Tbl3 Tcfap4 Tdp1 Telo2 Thap4 Thoc5 Ticam1 Timm13 Tjp2 Tkt Tmed10 Tmed3 Tmem109 Tmem11 Tmem111  59 335 13 426 59 79 380 88 20 327 23 50 85 92 82 11 80 15 29 11 50 12 18 24 18 47 10 231 29 326 16 79 67 46 11  54 285 4 359 64 50 281 61 6 218 30 37 69 83 80 5 60 25 13 9 30 9 16 15 19 47 10 206 30 245 21 66 49 39 5  118 290 11 415 48 78 157 63 17 344 49 74 79 122 174 13 75 21 26 18 56 8 15 31 22 32 18 299 23 280 18 59 100 44 12  46 231 7 108 39 30 228 7 16 123 19 55 31 101 56 3 63 20 24 12 17 5 6 16 7 26 11 60 27 109 16 38 55 17 6  7 169 2 15 27 33 73 0 9 131 14 14 2 51 29 0 0 6 2 5 2 3 3 0 3 18 0 75 6 31 3 27 16 14 2  Tmem179b Tmem186 Tmem192 Tmem214 Tmem33 Tmem80 Traf2 Traf7 Trappc6a Trappc6b Trdmt1 Triap1 Trim28 Trim37 Trpc4ap Tsfm Ttc4 Tubd1 Twsg1 Txlna Txndc9 Txnl4a U2af1l4 Ube2j1 Ube2m Ubox5 Ubxn8 Ufm1 Umps Uqcrc1 Uqcrc2 Urod Usp10 Usp30 Usp42  33 19 12 17 51 14 13 24 24 15 5 31 260 25 43 167 40 12 90 33 13 34 18 25 124 10 33 19 134 79 218 107 94 10 19  38 2 14 22 56 15 15 24 28 17 5 38 186 32 28 210 30 8 77 21 18 34 13 14 77 7 19 15 76 84 177 139 76 9 9  37 18 13 21 37 9 12 14 35 15 12 29 119 33 67 239 30 14 59 16 17 34 17 34 110 14 32 11 128 130 313 182 48 14 19  40 11 11 10 33 8 8 16 16 17 6 14 36 8 24 74 23 8 70 21 8 10 8 21 42 5 24 9 46 85 165 87 28 9 7  7 4 2 4 21 5 3 8 4 9 0 13 57 3 12 76 10 2 11 5 4 8 10 8 37 2 17 4 41 46 139 76 13 0 2  216  Vac14 Vars2 Vbp1 Vcp Vegfb Vegfc Vps29 Vps37b Vps45 Vps53 Wbp11 Wbp2 Wbp7 Wdr18 Wdr36 Wdr42a Wdr5 Wdr74 Wdr75 Wdr8 Wipf2 Wwp2 Xab2 Xbp1 Xrcc6 Yars Zc3h4 Zfp385a Zfp410 Zfp593 Zfp629 Zmym4 Zmynd19 Znhit2  24 12 255 177 33 18 103 14 7 9 346 41 17 97 19 48 21 128 75 14 24 27 23 40 70 73 38 81 99 23 27 19 17 56  23 12 161 129 11 17 149 13 9 11 428 22 12 73 15 38 15 117 67 11 19 15 25 60 81 75 26 94 93 21 18 8 13 74  26 14 258 301 31 15 153 21 12 9 325 52 13 84 14 41 18 115 86 10 17 28 16 33 129 55 47 114 61 22 23 13 25 108  14 9 190 86 17 15 71 8 6 7 155 45 10 27 10 40 7 67 25 7 15 21 18 53 34 21 23 36 79 8 18 9 4 30  6 3 53 57 9 2 42 9 3 5 140 23 6 9 2 24 4 28 10 2 7 4 5 22 18 20 7 21 35 9 11 4 4 30  217  Appendix I Cluster I genes and their expression in the five Tag-seq libraries (in tags per million).  Symbol 1110002B05Rik 1700017B05Rik 1810063B05Rik 2310005N01Rik 2310057M21Rik 2510002D24Rik 2510006D16Rik 2810403A07Rik 4930519N13Rik 4931406P16Rik 4933407C03Rik 9130011J15Rik 9930104L06Rik A230046K03Rik Abcg2 Abi2 Acbd3 Actn1 Adam12 Adamts9 Add3 Adss Afap1 AI428936 Alcam Alkbh1 Amotl1 Ankrd10 Ankrd13a  E10 DLK1+ 154 10 57 19 10 19 49 36 6 14 83 22 8 13 26 37 13 301 8 20 65 72 13 5 70 20 18 26 12  E12 DLK1+ 7 2 6 6 0 0 2 8 0 2 6 5 3 6 7 2 7 91 0 0 18 14 3 3 32 0 0 7 0  E14 DLK1+ 12 4 14 2 5 5 8 11 4 5 8 6 2 4 14 3 4 104 0 2 11 10 1 2 34 8 5 7 1  E16 DLK1+ 193 10 36 13 7 13 25 42 11 19 63 17 11 12 32 21 17 271 17 44 38 38 20 11 180 18 10 27 7  adult liver 12 0 0 0 0 2 3 5 0 3 4 2 0 0 0 0 0 10 0 0 7 6 0 0 8 0 0 3 0  Ankrd37 Ap4e1 Armc8 Arsk Ash1l AW549877 B3galnt2 B3gnt1 BC024479 Bcat2 Bcmo1 Bid Bivm Brwd2 C430048L16Rik Ccar1 Ccdc117 Ccnt2 Cd248 Cd2ap Cd9 Cdc123 Cdc7 Cdh11 Cdkn1b Cdkn1c Cdkn2aip Cdr2l Cebpa Cep120 Cfl2 Chic2 Chmp4b Cirbp Cited2  12 10 16 7 16 40 13 15 11 12 26 23 39 15 23 71 12 6 8 15 17 65 21 13 19 1138 20 5 162 11 98 23 32 56 131  5 0 0 0 4 10 3 2 0 5 2 6 3 3 3 4 0 4 0 3 6 4 7 0 5 148 6 0 27 3 12 0 4 25 34  2 3 2 3 3 22 4 7 2 2 10 7 4 3 6 12 3 0 0 5 16 20 1 0 7 329 5 1 15 0 16 1 6 17 31  19 14 7 11 13 107 8 15 9 11 16 16 24 13 18 41 6 11 11 20 47 33 14 22 35 663 13 12 251 7 58 17 14 61 84  2 2 2 0 0 8 0 2 0 0 0 0 4 3 0 2 2 0 0 2 2 4 0 0 6 0 2 0 43 2 9 0 4 7 11  218  Clic1 Cog3 Col5a1 Col5a2 Commd4 Crlf3 Ctdspl2 Ctnnal1 Cyb5r1 Cyfip1 D6Wsu116e Dact3 Dhx57 Dnajb1 Dpysl2 Dstn Dusp12 Dusp22 Dusp6 Dync1i2 Efnb1 Eid1 Eif2s3y Eif4e3 Eln Emilin1 ENSMUSG00000074460 Ep300 Epb4.1l3 Epha7 Eral1 Ergic2 Ewsr1 Ext1 F2r  15 11 13 18 25 49 9 48 19 123 10 21 21 18 26 13 67 16 98 113 23 95 10 7 10 34 20 16 11 10 16 20 155 11 35  2 3 0 0 0 20 0 8 2 22 3 3 13 3 12 2 26 4 19 28 8 17 2 0 6 6 5 2 2 0 3 0 71 3 13  6 2 0 0 7 14 0 23 5 27 5 1 9 5 5 0 28 6 14 39 10 25 4 0 5 5 4 2 0 1 5 2 30 2 15  11 9 27 36 11 48 10 36 12 118 15 12 41 15 32 14 76 25 201 118 28 59 15 17 29 17 12 10 6 11 13 13 138 13 43  0 2 0 0 3 5 0 4 0 10 2 0 0 0 5 0 2 4 20 8 2 3 0 2 0 0 2 2 0 0 3 2 4 3 7  Fabp5 Fam108b Fam116a Fam118a Fam176b Fam20a Fam49b Fbn1 Fbxo38 Fbxw5 Fbxw9 Fchsd2 Fem1b Fkbp1a Fmnl2 Foxn2 Fzd1 Gab1 Gadd45a Gapvd1 Gja1 Glg1 Gltp Gmfb Gnai3 Gnaq Gnb4 Gpc3 Gpi1 Gpkow Gpr177 Gpr180 Gpx8 Grb7 Gtf2a2  29 19 42 19 16 18 59 7 25 23 34 8 15 184 8 17 12 8 11 44 47 18 6 15 64 49 9 2103 161 21 17 22 13 56 65  13 6 11 2 0 0 11 0 6 6 11 3 2 21 1 2 2 3 3 8 4 0 0 2 19 6 2 690 27 3 0 0 0 8 15  15 4 11 3 9 6 9 1 10 13 9 5 4 41 0 1 5 2 1 12 4 5 3 5 10 7 2 337 27 9 5 2 2 8 18  45 13 27 12 25 9 50 18 32 25 52 15 7 87 12 14 25 15 20 31 35 22 12 9 54 41 10 2026 133 18 17 12 17 31 38  0 0 6 0 0 2 0 2 6 0 0 0 0 7 3 0 2 0 0 2 5 0 0 0 13 8 3 0 5 0 3 2 0 4 0  219  Gtf3c2 Guf1 H1f0 H2afv H2-K1 H47 Hccs Hes1 Hif1a Hmcn1 Hnrnph1 Hs3st3b1 Hsp90b1 Hspa13 Id1 Id2 Ier2 Ier5 Ifrd1 Igf1r Impad1 Isoc1 Itgb1 Itm2a Josd1 Kif15 Kif3a Kifap3 Klf7 Lamb1-1 Laptm4a Ldlrad3 Lgals1 Lmbr1 LOC100045315  97 15 18 113 11 106 31 88 34 7 102 40 187 62 43 473 26 11 40 14 11 9 158 314 6 23 11 27 4 80 210 31 60 16 10  19 4 6 6 0 9 8 10 8 0 19 9 34 10 4 90 3 0 19 3 1 0 33 32 0 9 4 4 0 10 23 6 16 2 3  19 3 4 30 1 26 7 10 3 0 3 9 20 13 10 75 7 0 8 2 4 4 84 66 3 7 3 4 2 3 41 7 21 5 1  49 15 21 66 18 70 19 180 24 13 53 35 126 47 37 398 53 15 71 17 9 15 161 605 14 20 12 14 13 76 224 18 100 12 11  10 2 2 3 0 14 0 0 5 0 10 4 7 5 4 36 3 2 7 0 0 0 20 0 2 0 0 0 0 5 17 3 0 0 3  LOC100045716 LOC100046895 LOC100047012 LOC675709 Lpl Lpp Lrig3 Lrrfip1 Lsm14a Lxn Magt1 Map3k1 Mapk14 Mapkap1 Matr3 Mbip Med13l Med6 Meg3 Mest Mettl1 Mfap2 Mmp14 Mrps18a Msn Mtdh Mtx2 Nampt Nav1 Nbea Ncapg2 Ncoa6 Ndn Ndufv1 Nes  27 28 14 21 85 9 4 18 72 12 5 11 39 23 204 18 13 34 258 237 27 36 12 80 8 78 94 23 10 15 26 11 14 98 6  0 2 0 6 39 0 0 2 13 0 5 0 6 3 53 2 4 10 47 7 0 0 0 8 3 12 11 5 0 2 4 0 0 25 0  3 3 2 4 30 4 2 6 10 3 1 1 16 11 86 2 2 7 103 12 2 1 4 19 2 19 19 4 2 5 5 2 0 35 2  39 17 9 15 111 11 13 27 30 18 17 9 28 18 170 18 15 21 792 387 20 56 10 45 19 44 66 23 5 15 19 6 14 78 12  0 0 0 0 0 0 0 0 7 0 0 0 2 4 22 0 2 2 0 3 0 0 0 0 0 3 11 0 0 0 0 0 0 13 0  220  Nfkb1 Nrp1 Ogdh Ogfrl1 Oxct1 Pan3 Parp2 Pdgfrb Pdia4 Pdk1 Pdk3 Pdzrn3 Phactr2 Phc2 Phlda3 Phtf1 Pigm Plaa Plekhf2 Pmp22 Pnn Pogz Ppp2r1a Ppp4r2 Prkar2a Prkcd Prpf18 Prpf40a Prpsap2 Prune Psmd10 Ptbp2 Ptpn14 Rab6 Rad51ap1  24 18 68 8 46 9 15 8 261 44 35 6 9 35 9 10 24 170 10 36 51 40 86 14 44 39 19 12 38 13 28 55 8 144 10  3 0 2 3 10 1 0 0 39 10 6 0 3 3 0 2 5 26 5 2 20 11 22 3 0 13 6 2 9 0 3 5 0 5 2  11 1 16 5 9 3 5 0 41 10 6 1 3 8 2 0 5 28 3 0 13 11 22 4 6 11 4 1 13 4 6 9 0 42 1  27 19 31 11 53 13 11 14 121 25 18 10 11 15 11 6 16 73 17 59 43 34 59 10 25 41 14 7 30 7 15 45 11 75 5  2 2 5 0 2 0 0 2 19 3 0 2 0 3 0 0 2 13 0 2 7 2 7 2 4 6 0 2 6 0 2 4 0 10 0  Rap2b Rarb Rbm5 Riok1 Rlf Rnaseh1 Rnf149 Rnf166 Rnf216 Rnf6 Rps6ka6 Scamp2 Sec61a2 Selk Sema4b Senp2 Sfrs12 Sfrs8 Sh3bgrl2 Sh3bgrl3 Sh3rf1 Shprh Slc1a4 Slc25a24 Slc30a7 Slc38a2 Slc43a3 Slc7a1 Smad1 Smarca1 Smndc1 Smu1 Snapc3 Snapin Snx25  6 18 35 23 13 45 37 19 18 25 62 28 46 26 7 34 16 37 10 22 7 10 15 12 13 31 14 20 54 28 10 78 22 19 18  0 8 8 7 2 11 11 8 0 0 13 6 13 6 2 4 2 10 0 3 3 0 6 0 0 2 2 3 11 8 2 8 3 5 4  2 4 12 3 2 11 14 8 0 6 16 17 22 12 5 3 6 14 2 6 0 1 5 6 3 11 3 3 8 5 4 21 3 4 7  11 23 39 19 10 37 53 27 12 26 39 48 45 59 11 19 20 38 8 14 10 14 14 14 11 28 39 13 40 19 11 40 11 16 23  0 0 9 2 0 2 10 2 0 4 0 3 4 7 0 3 4 4 0 2 0 2 0 0 2 4 3 0 9 0 0 5 2 3 2  221  Spnb2 Sppl3 Spred1 Srfbp1 Stard3nl Stk40 Taf2 Tbc1d1 Tceal8 Tcf12 Tgfbr3 Tgif1 Thoc7 Timp2 Tm9sf1 Tmem161b Tmem59 Tnfrsf21 Tnrc6a Tob2 Tomm34 Trip11 Tshz1 Ttc23 Tubb6 Ube2e2 Ube2g1 Ubqln2 Ubqln4 Unc119 Unk Usp45 Vamp3 Vamp4 Vasn  60 39 13 25 109 16 21 38 21 29 57 80 66 26 36 10 92 5 138 20 13 7 11 15 11 23 47 31 16 22 11 6 14 39 16  6 8 0 4 14 0 0 2 5 5 21 23 8 11 7 2 21 2 27 4 2 4 0 2 5 0 8 5 3 5 3 3 4 11 0  7 10 1 3 31 6 0 4 2 4 16 27 13 7 6 5 40 3 57 5 1 3 0 4 5 3 12 7 4 7 3 2 4 5 5  46 33 11 17 79 11 13 22 14 20 59 126 32 33 33 9 112 18 186 22 14 15 9 22 15 23 25 22 10 19 10 10 17 34 18  4 2 0 0 5 0 0 0 0 2 4 9 3 5 3 0 14 0 29 0 0 0 2 3 0 0 2 2 0 4 0 0 0 3 2  Vezf1 Vim Vps35 Wdr40a Wee1 Wipi1 Xrn2 Zcchc9 Zdhhc20 Zeb1 Zfhx3 Zfp263 Zfp275 Zfp292 Zfp36l1 Zfp710 Zfp821 Zfr Zmym2 Zmym5 Zmynd11 Znrd1  45 197 42 27 34 8 71 30 31 17 31 10 31 8 44 12 11 52 31 18 28 38  17 7 8 2 4 0 13 6 10 2 9 2 0 0 8 3 3 13 8 2 8 10  9 6 20 4 7 3 20 4 4 8 8 4 4 3 1 4 5 15 3 1 8 12  35 219 36 22 16 12 42 16 23 40 21 18 17 18 97 13 8 42 16 26 25 25  2 4 0 2 0 0 5 0 2 0 4 0 5 0 4 0 0 5 3 0 4 4  222  Appendix J Cluster J genes and their expression in the five Tag-seq libraries (in tags per million).  Symbol 0610007P14Rik 0610037D15Rik 0610037M15Rik 1110002N22Rik 1110018G07Rik 1700019H03Rik 1810055G02Rik 1810074P20Rik 2700062C07Rik 2700097O09Rik 2810008M24Rik 2900073G15Rik 3300001P08Rik 4930420K17Rik 4930547N16Rik 4930573I19Rik 5033414D02Rik 6430706D22Rik 9130227C08Rik A830080D01Rik Abca8b Abcf3 Abi3 Acaa1a Acat2 Acer3 Aco1 Aco2 Acot2  E10 DLK1+ 153 12 21 36 38 3 28 9 16 23 20 326 77 22 11 9 28 51 7 6 4 12 0 33 72 16 16 77 82  E12 DLK1+ 114 4 7 17 20 36 26 4 6 13 12 166 44 6 10 6 24 65 6 3 7 9 6 24 51 11 7 86 40  E14 DLK1+ 115 10 14 23 13 12 44 2 13 6 9 190 52 7 5 3 14 54 7 10 10 8 5 44 41 4 1 62 26  E16 DLK1+ 203 18 22 27 49 41 75 11 13 21 19 296 66 34 13 11 38 91 12 10 13 11 12 46 84 26 24 81 112  adult liver 14 0 4 7 17 3 8 3 2 2 7 86 9 10 0 2 3 8 2 0 0 0 0 11 12 4 7 13 18  Actb Actn4 Adcy6 Adpgk Adrbk1 Aff4 Agfg2 Agps AI314180 Akap7 Akr1b7 Akr1e1 Aldoc Alkbh3 Ambp Amdhd2 Ankrd46 Ano6 Ap1s1 Ap2a2 Apoa1bp Appl2 Arf2 Arfgef2 Arfip1 Arhgap1 Arhgef2 Arhgef5 Arih2 Arpc2 Arpc4 Arpc5 Arrb2 Asah1 Asph  977 79 23 36 17 23 25 29 66 8 7 26 5 17 21 12 47 14 30 25 81 44 33 15 23 134 22 5 49 152 97 131 8 74 12  399 9 13 20 4 22 17 22 13 7 5 13 6 5 24 2 14 7 8 14 34 46 23 8 8 120 10 7 21 21 36 51 2 83 10  450 46 12 16 18 14 12 20 58 4 22 26 10 16 16 8 31 3 20 17 33 34 29 14 16 92 12 9 22 73 52 57 5 54 7  1048 77 26 25 17 31 55 27 86 16 31 40 12 35 31 18 56 16 25 20 87 95 34 23 27 132 28 14 51 123 84 165 15 111 13  81 27 3 7 5 12 9 0 9 2 0 9 2 4 10 3 16 3 0 3 15 7 3 3 3 17 0 2 10 31 2 37 3 29 0  223  Atg16l1 Atp11b Atp1b1 Atp6ap2 Atp6v1a Auh Avpi1 B230219D22Rik B3gnt2 Baiap2 Baiap2l1 Baz2a BC005537 BC013529 BC020002 BC021381 BC037112 Bnip1 Brwd1 C030046E11Rik C130074G19Rik Cabin1 Cald1 Capza2 Cask Cbx4 Cc2d2a Ccl24 Ccnl1 Cd81 Cdr2 Cenpl Cept1 Chd2 Chd4  24 22 16 57 49 14 21 52 2 13 32 9 6 13 18 76 35 13 29 13 27 11 232 82 19 16 21 0 19 56 10 12 19 11 220  23 13 7 34 13 13 34 36 4 8 16 12 5 13 7 48 24 7 23 10 9 7 82 36 7 7 17 11 25 20 5 3 26 13 75  14 7 7 46 13 7 21 37 2 11 26 6 2 8 17 48 30 10 14 10 18 6 110 78 12 7 19 13 22 24 5 7 24 10 102  39 26 19 98 64 14 40 54 10 34 71 11 11 14 23 110 35 20 22 13 24 10 320 99 16 28 24 30 66 59 14 11 62 17 178  13 2 7 16 19 4 4 10 2 4 3 3 0 2 3 30 11 3 4 6 5 2 54 46 6 8 4 0 10 9 0 2 7 4 21  Chd6 Clk4 Cln5 Cln6 Clybl Cmas Cmpk1 Cmtm7 Cnn2 Cntln Copg Cox4i1 Cpne3 Cr1l Creld1 Cryz Csad Csk Csnk2a1 Ctage5 Ctf1 Ctsb Cxadr D19Wsu162e D330038O06Rik D930015E06Rik Dctn3 Ddx3y Ddx5 Dennd5b Derl1 Dffb Dhx34 Dido1 Dnajc18  11 19 27 5 13 54 22 11 15 7 137 717 70 14 13 24 9 14 38 49 9 115 41 34 17 9 70 17 95 28 31 11 10 17 16  7 37 12 3 3 26 21 7 14 13 125 359 56 8 6 18 2 8 21 81 6 113 47 12 23 2 30 13 34 15 17 8 8 9 6  3 12 14 8 13 29 10 6 29 7 141 562 58 17 7 21 7 6 10 37 0 171 43 18 20 3 51 23 52 17 29 4 5 5 10  14 69 21 10 12 36 20 13 50 23 170 559 128 22 9 39 17 12 35 113 10 338 80 63 29 11 60 24 106 21 44 10 16 16 12  0 13 4 0 0 8 7 2 0 2 7 87 23 0 2 7 7 0 5 32 0 65 23 21 2 3 15 0 11 8 21 0 0 2 3  224  Dnajc5 Dnttip1 Dopey2 Dpp8 E130112L23Rik E2f2 E430002G05Rik E430025E21Rik Eaf1 Eif2ak4 Emilin2 Emp2 Enc1 Epm2aip1 Eps15l1 Ermp1 Esd Esf1 Ets2 Evi5 Exoc3 Exoc5 Fam114a1 Fam126a Fam164a Fam168a Fam173a Fam173b Fam18b Fam3c Fam62b Fam76b Fasn Fbxo18 Fbxo33  12 43 14 35 27 12 9 37 47 25 4 15 13 53 64 15 265 44 45 28 10 90 9 46 11 16 35 13 93 23 25 14 116 25 13  4 30 12 26 36 4 11 20 20 19 12 18 3 50 30 6 202 33 29 40 5 87 8 32 6 18 26 4 32 6 59 16 98 18 7  11 18 20 22 27 9 10 18 20 17 18 11 8 45 36 13 220 26 33 25 5 36 16 55 12 3 24 13 53 13 28 7 61 12 10  15 46 45 55 36 11 19 33 36 27 37 17 13 113 62 13 281 36 57 73 7 86 16 55 18 18 35 13 93 22 74 22 127 35 11  2 11 4 4 14 0 2 6 7 3 4 0 2 8 6 4 95 4 8 18 2 32 3 10 0 7 8 3 9 0 20 2 53 15 4  Fbxw2 Fdxr Fermt2 Fez2 Fgfr3 Fkbp14 Fkbp9 Flna Fmr1 Fn1 Fnbp4 Fndc3b Foxa3 Frs3 Fubp1 Galc Gdi1 Gga2 Ggcx Gipc1 Gls Glt25d1 Glud1 Gmeb2 Gmppa Gna13 Gnpnat1 Gpbp1l1 Gramd1a Grb2 Gripap1 Gss Gstm7 Gsto1 Gtf2h1  23 22 51 12 12 16 75 66 62 360 27 48 62 14 27 5 84 171 50 11 31 51 1742 4 11 20 55 11 17 34 13 26 13 198 20  34 26 41 8 5 24 40 38 119 471 23 42 36 8 34 4 70 75 68 5 18 72 914 3 6 13 26 8 22 27 6 9 25 120 10  20 24 32 13 4 30 42 59 77 254 21 17 40 3 37 1 60 92 72 8 2 59 1710 5 8 10 53 0 14 14 8 20 26 173 12  30 33 128 24 14 38 112 86 114 673 38 89 58 12 52 11 119 201 126 9 29 79 2186 10 12 33 61 11 26 44 14 30 52 196 15  9 3 32 2 2 7 0 20 19 47 14 5 11 2 9 2 31 13 17 0 0 22 298 0 4 4 31 3 2 2 6 2 13 56 5  225  Gtpbp2 H19 H3f3b Hes6 Hist3h2a Hltf Hmg20a Hook3 Hps4 Hsd17b7 Idh2 Idh3b Ier3ip1 Ifngr1 Ift52 Igsf8 Ing5 Itgav Itm2c Jmjd1b Jmjd3 Jmjd8 Jund Kank2 Kctd3 Kctd6 Kdelc2 Kif21b Klf12 Klf2 Klhl13 Klhl20 Klhl23 Klhl7 Krcc1  7 6668 256 5 11 14 26 5 7 9 399 82 62 18 12 21 57 76 118 51 15 38 23 27 37 6 26 8 8 22 25 8 21 40 49  14 4286 119 3 10 19 17 1 5 9 381 68 28 19 4 7 21 45 95 56 14 15 14 23 17 5 11 6 6 10 15 5 17 23 13  8 9874 131 4 13 18 18 6 11 6 678 38 30 14 5 16 26 47 129 17 12 32 13 20 27 6 22 6 9 9 14 5 12 27 33  32 9547 370 12 24 44 22 11 12 20 656 94 63 44 10 20 56 132 158 47 18 41 26 24 38 17 24 10 13 26 35 12 16 49 57  7 0 23 2 0 14 7 0 0 4 110 15 11 10 2 4 4 29 6 16 8 2 3 6 3 3 4 0 2 6 5 0 0 7 6  Krit1 Ktn1 Lamc1 Lancl1 Large Lifr Litaf Lmo7 Lnx2 LOC100045448 LOC100045522 LOC100045542 LOC100046855 LOC100046998 LOC100047539 LOC100047794 LOC100048397 LOC677213 Lrba Lrrc25 Lrrc32 Lrrc8a Luc7l M6prbp1 Maf1 Maff Man1a2 Man1c1 Map2k3 Map2k6 Map3k2 Map4k5 Mapkapk2 Matn2 Mcoln1  15 20 23 27 22 40 86 21 5 26 28 50 39 8 17 102 12 53 11 0 2 13 65 68 38 4 20 19 19 18 7 45 51 5 14  25 9 29 18 7 62 37 28 0 11 21 36 12 4 5 63 10 34 2 8 16 13 58 49 19 3 10 18 12 12 10 20 33 27 3  15 6 5 17 13 18 56 24 5 15 16 27 22 7 16 26 8 37 7 10 10 7 35 42 28 5 12 12 14 13 5 20 53 55 8  28 18 35 24 19 88 205 36 11 41 46 39 44 13 17 92 19 80 7 19 28 21 57 71 30 14 17 19 23 16 18 56 69 87 11  5 0 4 4 6 26 33 5 3 5 2 4 7 2 2 24 2 32 2 3 3 4 7 20 4 0 5 0 5 0 0 10 33 2 0  226  Mcts1 Mdh1 Med25 Meis1 Mga Mgea5 Mgst2 Mmgt2 Mobkl2a Mon2 Mpp6 Mpzl2 Mtif2 Mum1 Mut Mxd1 Mycbp2 Myo6 Ncrna00153 Ncstn Ndrg1 Ndst1 Ndufa3 Necap2 Nfxl1 Nfyb Nme7 Npc1 Nt5c2 Ociad2 Ocln Ogt Orc3l Orc5l Otud4  19 612 58 27 44 33 14 27 13 11 38 26 10 57 11 6 29 3 14 18 83 12 84 5 28 114 24 33 20 9 6 43 25 17 66  23 300 53 16 38 35 9 22 12 0 15 59 5 23 5 7 22 6 6 5 45 2 26 4 21 44 27 61 8 8 6 27 16 6 47  10 398 59 20 24 19 25 14 15 5 21 26 5 24 8 5 23 7 7 12 63 9 60 4 8 63 14 35 13 16 9 15 13 8 39  24 658 69 27 69 51 30 23 16 12 28 57 19 48 17 14 39 14 15 19 152 16 65 10 33 135 26 58 18 23 11 64 22 14 53  10 247 5 4 11 4 0 2 0 3 5 7 3 2 4 0 7 2 4 0 17 3 19 2 9 21 5 17 6 2 0 17 2 2 9  Pard3 Pbrm1 Pck2 Pdcd6 Pddc1 Pde4dip Peli1 Pepd Per2 Pex11b Pex7 Pfkp Phlppl Pias1 Pigo Pigy Pitpnc1 Pitpnm2 Plekha2 Plxna3 Pnliprp1 Ppapdc1b Ppic Ppp1r12c Ppp1r9a Pppde1 Prepl Prpf38b Prpf39 Prss8 Pskh1 Ptpmt1 Ptprk Pygb Qk  22 198 25 30 71 28 25 31 15 5 30 5 16 14 16 10 16 10 6 9 12 30 67 45 38 12 6 102 15 6 34 58 13 14 37  11 90 20 5 20 28 6 19 32 2 27 2 10 10 6 2 8 8 4 7 11 19 49 33 57 9 11 69 13 8 22 23 10 11 25  20 85 33 22 35 24 7 33 15 5 21 8 10 12 10 7 12 7 10 6 12 20 58 51 33 7 5 51 8 11 32 42 11 16 18  42 173 32 31 53 33 33 46 34 12 40 12 21 22 17 11 29 20 19 11 14 32 87 59 71 12 10 98 20 24 32 56 22 17 36  13 24 0 12 12 7 11 6 0 0 18 0 2 7 4 2 11 7 2 0 0 4 8 23 6 0 3 13 2 0 5 18 7 0 7  227  Rab11a Rab21 Rab24 Rab3il1 Rab4a Rab5c Rabgef1 Ralb Ranbp9 Rasa1 Rbm25 Rbp2 Rbpms2 Rc3h2 Reep4 Reep5 Rgl1 Rhob Rhobtb1 Rhoc Rhoq Rnd3 Rnpepl1 Rock2 Rogdi RP23-195K8.6 Rp2h Rragc Rrbp1 Rsbn1l Rspry1 Safb2 Sar1a Scd2 Scfd1  135 9 34 4 8 70 10 16 15 18 86 12 17 3 46 7 11 56 8 31 14 21 35 23 33 8 12 35 287 7 20 23 89 523 37  94 5 36 7 8 36 7 6 6 4 78 18 12 6 13 6 9 48 11 38 10 11 11 17 31 2 11 24 219 4 8 21 19 338 29  124 2 29 5 21 28 10 8 5 21 32 31 15 3 31 5 18 70 8 40 5 7 42 10 41 8 5 26 497 3 9 20 52 311 42  150 13 40 10 27 61 18 18 19 28 101 66 21 10 42 14 23 112 12 49 16 26 58 35 49 19 16 43 495 10 18 32 65 654 58  10 3 16 0 5 15 6 0 6 3 17 0 9 0 3 2 5 20 5 5 0 4 6 2 0 2 3 11 87 2 0 15 13 0 3  Sdc3 Sdha Sec23a Sec24d Sema4a Serpina1b Sesn3 Sfrs14 Sgcb Sh3bgrl Sh3glb2 Siah2 Sik1 Sirpa Ski Skiv2l Slc20a1 Slc26a11 Slc36a1 Slc43a2 Smarcal1 Smc1a Smc6 Snx1 Soat1 Son Sorbs3 Sp1 Sp4 Spag5 Spen Spink3 Sptlc2 Srf Srp19  11 130 30 7 29 16 14 20 7 154 10 14 58 29 95 13 29 8 18 3 18 40 22 34 46 125 8 19 22 29 14 58 13 21 91  13 74 35 4 15 4 17 18 4 126 0 9 31 15 53 10 31 9 33 9 16 19 7 18 21 55 6 7 13 24 6 52 12 31 44  19 131 27 5 23 11 6 15 6 67 7 15 16 10 62 8 32 5 34 9 11 34 7 14 22 62 7 9 22 30 8 116 12 24 55  41 154 58 13 47 28 16 36 14 210 10 17 82 39 87 16 73 13 59 19 16 44 23 30 63 142 10 17 28 31 15 173 14 43 68  6 49 9 4 20 10 2 4 4 62 0 0 14 14 24 5 3 3 8 2 3 4 7 4 2 52 4 3 6 0 3 0 4 10 4  228  Srp72 Srpk2 Stard4 Stil Stk11 Stk3 Stradb Stx12 Stx7 Swap70 Syf2 Tanc2 Taok1 Tax1bp1 Tbc1d8 Tbrg1 Tceal5 Tceb3 Tcf19 Tfrc Thpo Tial1 Tm2d1 Tm9sf2 Tm9sf4 Tmcc3 Tmem176b Tmem194 Tmem37 Tmem39a Tmem64 Tmem65 Tnfaip2 Top2b Tor1b  95 23 18 8 24 24 14 45 28 9 21 11 37 44 3 46 22 30 17 109 6 138 72 24 29 25 65 10 79 124 44 20 13 156 11  49 15 12 10 6 6 13 23 24 8 7 6 25 49 7 14 50 14 8 69 4 89 38 14 19 12 43 13 108 62 45 13 8 87 5  43 8 17 2 16 11 20 23 16 4 17 5 17 38 9 31 63 24 14 39 6 77 56 18 20 8 35 12 123 99 25 15 7 76 5  81 28 46 10 22 35 22 42 26 10 21 9 46 65 19 44 87 35 17 97 11 178 68 25 24 22 133 18 173 113 48 26 19 112 10  12 10 2 0 2 8 4 4 7 4 6 0 15 21 2 6 0 9 0 7 0 36 18 7 4 2 20 0 9 2 15 0 4 32 2  Tpp1 Tpp2 Tra2a Trabd Trappc2l Trim24 Trim26 Trit1 Trpm7 Tsc1 Tsc22d2 Ttc38 Ttf1 Ttyh2 Twf1 Txndc4 Ubac1 Ube2a Ube2d3 Ube2g2 Ube4a Ubn2 Ubr1 Ufd1l Uhrf1bp1l Upf2 Usp3 Usp40 Usp52 Utrn Vamp2 Vcl Vcpip1 Vezt Vta1  38 59 203 14 11 21 12 10 20 10 27 18 12 3 47 51 35 17 160 35 17 22 11 16 14 17 11 15 60 11 6 34 9 12 13  36 48 175 2 11 18 7 12 14 9 18 12 6 2 20 30 13 13 88 10 14 33 5 9 7 20 6 9 82 4 8 10 15 13 9  18 32 118 12 10 8 10 15 4 12 15 12 11 9 25 35 14 13 132 20 6 24 3 4 9 9 10 14 29 13 9 32 4 13 4  50 64 217 13 15 25 15 19 28 18 29 34 16 14 36 49 30 19 146 31 17 37 15 20 17 22 11 16 90 13 21 67 14 17 11  19 14 17 2 3 7 4 2 11 2 6 8 4 2 10 14 7 2 50 7 3 11 4 0 3 0 0 3 24 4 2 3 3 6 0  229  Vwa5a Wapal Wbp4 Wdr1 Wdr32 Wdr48 Wsb2 Yipf1 Zbed3 Zcchc11 Zfp191 Zfp26 Zfp260 Zfp395 Zfp507 Zfp518 Zfp644 Zfp653 Zfp74 Zkscan3 Zmpste24 Zmym3 Zswim6  5 9 29 56 15 13 15 25 22 49 19 10 27 8 23 24 26 9 11 33 43 15 3  5 11 31 64 17 5 8 17 16 42 6 8 15 2 23 15 9 11 7 17 26 12 6  5 7 21 59 8 8 6 28 13 49 14 3 16 5 13 15 11 4 5 7 31 12 3  12 10 29 96 18 11 16 28 19 75 14 13 21 16 24 28 23 16 10 36 35 18 10  0 4 3 7 6 0 2 9 3 28 3 0 8 5 2 5 4 2 3 4 8 0 0  230  Appendix K Cluster K genes and their expression in the five Tag-seq libraries (in tags per million).  Symbol 1300014I06Rik 1300017J02Rik 1810011O10Rik 2310016C08Rik 2810055F11Rik 3110043O21Rik 4930402H24Rik Abcd2 Abcg1 Acat1 Acot12 Acox2 Acsl5 Adamts1 Adamts15 Adfp Adra1b Adrbk2 Afm Agpat9 Ahr AI182371 Alb Amacr Ankhd1 Anxa1 Apbb1ip Apcs Apoe  E10 DLK1+ 15 3 3 50 5 6 5 2 0 25 0 0 47 16 2 35 3 14 90 0 0 0 5426 13 3 0 0 6 919  E12 DLK1+ 11 6 0 51 5 3 2 8 4 44 3 9 32 1 0 3 31 16 72 0 3 0 48824 14 3 0 2 18 445  E14 DLK1+ 35 46 0 46 8 7 3 7 2 114 5 38 129 7 2 16 47 5 135 3 5 4 47360 9 3 2 6 46 789  E16 DLK1+ 57 151 50 366 18 19 15 20 14 222 23 60 195 115 11 68 119 51 236 12 31 13 108448 46 13 17 13 105 1856  adult liver 27 105 12 4 12 4 2 19 4 75 23 29 99 52 0 46 89 0 182 7 8 9 47907 12 5 2 4 104 1586  Apoh App Arap2 Arhgap29 Arhgdib Arrdc3 Arrdc4 Ass1 Axl Axud1 B230380D07Rik BC048546 Bcl3 Bgn Bhlhe40 Bhmt2 Birc3 Bok Bphl C1qa C5ar1 C77080 C8b Car2 Ccbe1 Ccdc80 Ccl6 Ccl9 Ccng2 Ccnl2 Cd36 Cd38 Cd48 Cd53 Cd93  6 18 1 4 6 20 8 47 8 3 7 0 5 4 3 12 2 2 32 0 0 7 2 5 6 4 0 2 10 27 0 4 0 1 0  346 5 0 1 11 10 1 55 2 2 5 1 3 0 0 11 3 0 29 2 2 11 51 10 0 0 4 5 5 29 0 0 0 13 0  529 6 1 3 10 7 4 49 5 0 2 7 5 1 4 31 5 1 34 2 0 32 122 14 1 1 4 16 5 20 1 12 3 13 2  1208 80 14 11 29 128 24 168 46 24 26 16 11 27 50 83 12 19 77 19 10 42 155 35 29 12 15 62 30 69 11 41 21 71 12  1025 11 3 9 12 12 13 151 2 0 3 7 5 9 6 27 0 11 50 4 0 34 98 11 5 4 0 30 14 51 4 14 2 4 0  231  Cdh5 Cebpd Chd3 Clec4a1 Clec4n Clec7a Col14a1 Col4a1 Col4a2 Colec10 Cp Crip1 Crip2 Csf2ra Csf2rb Ctdspl Ctsc Ctss Cx3cr1 Cxcl12 Cxcr4 Cybb Cyp2d22 Cyp3a16 Cyp4f13 Cytip Dak Dcn Dcxr Dgat2 Dgkq Dnajb9 Dqx1 Dse Ech1  1 5 10 0 0 0 0 41 5 6 0 2 5 15 0 7 4 0 0 8 0 0 4 0 3 0 13 4 3 186 4 49 12 14 28  0 2 16 4 4 8 0 19 2 0 2 2 4 16 4 5 27 10 4 6 0 5 4 0 4 0 2 2 4 132 3 2 41 8 34  1 3 12 2 7 8 0 20 2 1 19 5 2 15 6 5 24 21 4 7 2 6 11 4 9 1 16 0 4 195 1 28 33 8 52  31 29 51 16 10 42 67 98 22 57 139 13 26 56 26 27 91 182 14 63 16 53 22 13 17 11 50 169 10 500 11 100 138 59 118  4 0 7 0 4 27 13 77 5 9 84 0 18 6 5 5 76 10 0 4 0 18 16 6 10 0 37 59 9 181 4 42 0 6 47  Echdc3 Ecm1 Ednra EG624219 Ell2 Emr1 Enpp2 Entpd1 Epb4.2 Errfi1 Es22 F5 Fabp4 Fah Fam105a Fam126b Fam84a Fcer1g Fcgr1 Fcgr3 Fga Fgb Fgd6 Fli1 Fuca2 Fyb G0s2 Gadd45g Gas6 Gatm Gdf15 Gdf2 Gpm6a Gpm6b Grn  5 7 14 33 17 0 0 0 0 70 0 18 0 57 1 3 1 0 0 1 177 565 5 2 7 0 31 8 3 2 2 10 3 4 20  4 0 3 120 15 19 7 8 3 39 1 56 1 61 2 4 0 4 4 4 160 339 3 2 12 6 23 2 1 5 4 0 4 0 21  10 0 1 99 22 18 5 2 1 62 0 100 5 173 0 4 0 4 5 19 329 500 8 4 4 4 80 1 1 6 4 1 2 1 33  16 14 45 377 59 47 34 21 16 729 13 163 18 340 12 21 10 18 14 33 1650 1142 16 16 22 20 137 28 12 45 25 35 15 13 71  10 13 8 172 14 17 11 0 0 233 12 125 10 176 0 12 2 5 0 16 601 1010 8 0 11 3 62 10 9 19 0 4 3 4 32  232  Gucy1b3 Hao3 Hck Hebp1 Hfe Hgfac Hif3a Hist1h1c Hist1h2bc Hmgcr Hsd3b3 Icam1 Id3 Ifi30 Igfbp1 Igfbp3 Il15ra Il18bp Il1a Inpp4a Insig1 Irf2 Irf8 Islr Itga8 Itih3 Itpripl2 Kat2b Kcnq1ot1 Kctd12 Kdelr3 Kel Klf10 Klhl24 Lamp1  9 0 0 32 2 4 5 10 2 74 0 9 86 7 50 16 0 0 0 9 4 27 3 5 18 8 3 5 1 3 1 0 14 3 43  20 5 2 70 7 7 12 14 0 107 0 10 10 29 86 2 0 14 4 36 3 24 5 0 0 82 7 0 1 32 1 2 12 0 15  18 13 5 96 20 9 6 21 4 102 3 13 14 38 125 1 3 16 4 50 1 22 5 0 3 88 9 0 4 26 1 0 13 4 30  63 42 11 219 29 35 43 102 17 374 11 66 196 99 475 50 16 27 85 93 60 58 22 21 151 251 22 16 12 68 10 15 104 25 111  4 21 3 210 18 2 0 22 10 13 9 3 61 19 209 9 12 16 15 37 14 45 0 4 17 37 5 5 0 49 0 0 4 11 79  Laptm5 Lgmn Lhfp Lilrb4 LOC100045753 LOC100047749 Lrat Lst1 Ly86 Lyve1 Mafb Mafk Mal2 Malat1 Man1a Map3k8 Mertk Metapl1 Mgst3 Mlxipl Mpeg1 Mt2 Mxra8 Myd116 Mylip Mylk Nfkbia Nfkbid Nfkbiz Ngfr Nid1 Nlrp6 Npal1 Nr1h4 Nr3c1  1 21 5 0 7 5 0 1 1 0 3 6 4 5 15 0 3 2 30 22 0 2 5 7 0 4 6 1 3 5 37 0 0 0 18  5 19 0 4 0 13 0 7 4 0 4 4 3 5 20 2 0 3 15 84 6 10 0 4 0 23 5 2 2 0 48 5 0 0 8  17 8 0 0 4 25 0 8 6 7 10 2 9 1 37 0 5 3 27 49 5 25 2 3 0 22 3 1 3 0 94 4 6 3 18  28 95 19 19 18 41 35 42 34 56 26 28 25 51 63 13 11 10 151 159 46 100 26 38 10 67 79 11 28 34 286 26 37 21 76  7 25 7 2 5 33 20 0 21 3 26 0 9 11 31 0 0 2 17 133 7 0 11 4 5 14 12 0 0 10 15 21 4 14 21  233  Nr4a1 Nrbp2 Nrn1 Obfc2a Osgin1 Ostf1 OTTMUSG00000000971 Pde4c Per1 Pilrb2 Pim3 Pip4k2a Pld4 Plekha6 Plk3 Plxnc1 Pnrc1 Prelp Prex1 Ptgs1 Pygl Rac2 Rdh5 Rgs2 Rhoj Rhou Rras Rspo3 Rtp3 S100a16 Saa3 Saa4 Scarb2 Sdc4 Sema6d  1 0 0 2 28 11 0 0 6 0 13 3 0 2 2 4 48 0 3 4 25 2 2 3 1 38 8 4 1 6 0 14 10 34 5  0 4 8 2 87 5 0 0 13 0 6 2 0 4 0 0 17 0 0 15 54 16 0 1 3 25 4 0 4 0 1 34 12 5 0  0 4 44 0 99 7 1 1 5 4 4 10 3 5 1 1 32 0 3 17 69 23 6 1 4 32 11 0 10 3 7 49 23 12 5  12 22 131 13 292 31 12 13 26 11 45 22 15 17 41 15 98 18 12 66 120 33 10 22 10 97 31 16 20 16 14 77 44 90 27  2 10 57 0 56 10 5 3 15 5 20 0 0 10 5 7 37 16 0 0 54 16 3 3 0 42 13 3 0 3 0 47 15 23 2  Serinc3 Serinc5 Serpinb6a Sgk1 Sgpp1 Slc17a3 Slc17a5 Slc23a2 Slc25a25 Slc25a37 Slc36a2 Slc37a4 Slc40a1 Smoc2 Snhg11 Socs2 Sod3 Sorbs2 Spic Stard8 Stk17b Sult1a1 Syne2 Tcp11l2 Tff3 Thbd Tie1 Timd4 Tiparp Tm4sf4 Tmem204 Tmem25 Tnfrsf11b Tnfrsf1b Tnnc1  85 14 15 13 5 0 6 7 13 20 8 8 9 2 3 7 1 3 0 3 4 0 8 1 3 1 1 1 15 20 0 4 2 2 6  82 11 7 21 2 14 4 5 14 15 46 4 25 0 0 0 0 5 2 7 2 0 0 0 2 0 0 9 12 10 0 3 0 8 17  122 29 13 8 1 23 4 5 12 25 79 12 35 0 3 0 0 5 2 4 3 14 3 0 5 4 3 6 6 25 2 14 0 10 28  295 71 46 49 14 48 13 23 69 102 290 27 146 14 11 20 37 22 14 16 17 280 16 10 21 12 12 19 62 59 16 37 13 19 65  177 7 9 12 4 46 7 11 38 5 0 16 13 2 0 10 37 2 0 4 5 276 13 8 0 9 3 6 2 54 9 11 2 11 0  234  Trim2 Trim30 Trp53inp1 Ttc36 Ttc39c Ttr Tuft1 Tyrobp Ulk1 Xist Xkr9 Ypel5 Zeb2 Zfp655  4 0 2 3 6 66 5 1 10 1 0 9 28 8  2 4 1 5 6 20 4 8 13 2 0 15 6 0  5 1 1 43 6 47 6 13 12 0 0 14 17 7  15 13 24 83 14 132 18 58 39 11 14 50 100 24  0 10 10 44 10 74 0 8 10 6 5 29 13 12  235  Appendix L Cluster L genes and their expression in the five Tag-seq libraries (in tags per million).  Symbol 2310043N10Rik 8430408G22Rik 9030425E11Rik Acta2 Adra2b Adrb2 Akap12 Amotl2 Atf3 Btg2 Capn6 Cav1 Ccdc3 Ccl2 Ccl3 Ccl4 Ccrl2 Cebpb Col11a1 Col1a1 Col1a2 Col3a1 Col6a1 Col6a2 Creb3l1 Ctsk Cxcl1 Cxcl10 Cxcl16  E10 DLK1+ 6 1 3 4 2 5 2 14 17 18 30 1 30 0 0 0 1 2 3 43 124 369 11 7 4 0 0 1 0  E12 DLK1+ 0 2 0 0 0 0 0 0 4 2 0 0 1 0 0 0 0 0 0 0 4 0 2 1 0 0 0 0 0  E14 DLK1+ 3 11 1 1 1 2 0 4 0 8 0 0 3 0 1 1 1 0 0 0 2 2 1 1 0 0 1 1 0  E16 DLK1+ 126 489 12 42 11 16 19 39 104 199 90 15 75 28 14 49 28 16 28 244 472 1486 82 39 10 20 85 47 12  adult liver 6 14 0 0 0 0 0 0 4 4 0 0 2 0 0 0 0 0 0 0 4 17 8 0 0 0 0 0 0  Cxcl2 Cyr61 Ddx26b Dusp1 Dusp2 Eef2k Efna1 Egr1 Egr2 Emp1 Emr4 Fbln5 Fhl2 Fos Fosb Fstl1 Gadd45b Gdf10 Gem Ggt5 Gucy1a3 Hba-a1 Hba-a2 Hbegf Hhip Hist1h1d Hspa1b Il1b Irf1 Jdp2 Jun Junb Klf6 Lpar4 Lum  0 50 2 16 1 5 2 206 3 1 0 0 6 96 9 104 0 16 3 0 1 0 0 3 0 1 4 0 5 2 140 9 18 9 82  0 4 0 0 0 0 0 47 0 0 0 0 2 0 0 6 0 0 0 0 0 3 0 0 0 0 1 0 2 0 5 0 1 0 5  0 4 0 0 0 2 0 65 0 0 0 1 2 1 0 4 0 1 0 1 2 1 0 0 1 0 0 0 8 1 6 0 1 0 9  111 145 32 97 23 18 20 1389 10 33 13 39 35 496 37 243 17 102 17 15 19 35 24 10 15 29 39 28 99 19 350 65 79 21 519  0 5 0 7 0 0 0 10 0 0 0 2 0 0 0 10 0 3 0 0 0 0 0 0 0 0 0 0 2 0 2 0 0 0 16  236  Mmp2 Mmp23 Morc4 Olfml3 Pde1a Plagl1 Postn Prg2 Prtn3 Rgs1 Rgs16 Robo2 S100a8 S100a9 S1pr3 Scarf2 Serpine1 Sfrp1 Slc4a1 Socs3 Sparc Sparcl1 Srpx Stfa3 Sulf1 Tagln Tmem45a Tnfaip3 Wisp1 Zcchc5 Zfp36  42 2 1 3 2 64 32 0 2 0 1 8 2 0 6 5 3 45 0 16 141 0 2 0 4 4 5 0 5 0 1  2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 2 16 0 0 0 0 0 0 0 0 0 0  2 0 2 0 0 0 1 2 2 0 0 0 8 2 0 0 0 1 2 0 14 0 0 0 0 0 0 0 1 0 0  163 12 17 12 13 173 91 28 20 44 18 23 122 59 15 17 34 138 35 62 552 30 15 15 11 14 16 27 15 10 47  4 0 0 0 0 0 5 0 0 0 0 0 0 0 0 0 0 0 0 3 13 0 0 0 0 0 0 0 0 0 0  237  Appendix M Cluster M genes and their expression in the five Tag-seq libraries (in tags per million).  Symbol 0610012D14Rik 1110028A07Rik 1190003J15Rik 1300002K09Rik 1300018J18Rik 1810014F10Rik 1810020D17Rik 1810046J19Rik 2010011I20Rik 2310005N03Rik 2310007H09Rik 2310030G06Rik 2310051E17Rik 2410015M20Rik 2610507B11Rik 3110049J23Rik 4933439F18Rik 5830404H04Rik Aadacl1 Aadat Abcb6 Abhd14b Acad10 Acad11 Acad8 Acot4 Acox1 Acp6 Acsl1  E10 DLK1+ 3 4 3 1 4 4 10 55 7 11 5 2 4 42 5 2 43 4 0 13 27 24 3 7 35 17 46 14 106  E12 DLK1+ 11 0 8 17 5 14 6 44 2 11 7 0 0 41 3 3 8 2 0 9 15 15 12 7 98 18 102 2 140  E14 DLK1+ 12 1 22 10 5 14 13 37 5 8 7 1 0 56 5 12 12 2 4 6 15 23 9 8 85 28 244 5 111  E16 DLK1+ 14 0 39 24 10 8 8 25 12 6 2 2 12 28 2 30 15 3 2 21 8 42 7 22 34 20 355 17 346  adult liver 41 12 79 32 44 28 54 138 31 27 29 10 39 132 10 69 65 27 12 51 47 78 40 37 189 89 870 23 1317  Acss2 Adam9 Adck2 Adhfe1 Adipor2 Adk Afmid Agt Agtpbp1 Agxt Ahcy Ahsg AI464131 AI661453 Aifm2 Ak2 Ak3 Akap11 Aldh3a2 Aldh5a1 Aldh6a1 Aldh8a1 Aldob Alg9 Alkbh5 Als2cl Ank Antxr2 Apoa2 Apob Arg1 Arhgef19 Arl8a Asgr2 Asl  2 16 2 0 36 197 9 2 9 1 32 17 2 0 1 11 16 7 55 8 31 0 17 18 66 8 5 11 2125 191 17 1 8 32 108  7 2 4 0 31 236 14 16 6 3 14 373 4 0 2 5 10 4 104 16 18 4 153 22 63 10 6 90 2374 298 10 3 5 42 316  19 5 5 5 33 340 30 28 8 12 19 971 8 3 3 22 22 6 162 13 37 6 608 19 68 7 2 36 2702 142 12 1 7 98 253  18 10 2 14 36 258 27 33 19 14 22 1457 10 1 4 6 27 7 93 6 122 8 754 6 55 5 1 14 2961 492 252 1 14 35 83  43 21 14 41 374 610 89 66 25 67 133 4016 16 12 14 74 51 19 680 46 158 13 1927 76 165 20 22 145 9304 1908 611 14 16 222 776  238  Atf5 Atf6 Atg4b Atp6v0d1 Atp8b1 Atrn Atxn7 Avpr1a B2m Bat5 Bbox1 BC006779 Bckdha Bcl2l1 Bcr Bfar Blcap Brap Bst2 Btd C1galt1c1 C1r C1rl C330011K17Rik C530044N13Rik C630028N24Rik Cabc1 Calcrl Caml Car3 Cat Cbr1 Ccbl2 Ccs Cda  29 14 8 12 2 8 2 0 176 15 0 6 16 3 7 71 4 7 6 1 14 1 8 6 11 9 2 10 18 2 60 40 14 22 4  183 11 0 4 11 14 4 0 501 15 3 2 12 4 21 181 3 6 2 6 9 11 41 5 13 15 21 0 16 52 104 18 16 18 30  114 13 4 23 13 6 2 6 322 32 7 6 36 4 12 40 5 7 5 6 6 15 21 8 35 17 48 2 8 197 130 34 20 58 17  45 9 1 14 23 9 10 7 510 15 15 12 18 4 1 69 4 5 3 3 7 17 17 6 13 22 25 18 8 711 98 17 27 39 3  402 39 14 33 24 50 14 13 3091 60 17 41 61 14 61 538 20 31 29 16 25 37 95 26 51 67 112 23 68 1136 1195 76 57 145 105  Cdc37l1 Cdc42bpb Ces6 Cflar Cgnl1 Chac2 Chchd10 Chka Chn2 Chuk Cisd1 Cish Cml1 Cmtm6 Commd3 Cpn2 Cpt1a Crb3 Creb3l3 Creg1 Crot Crp Cryl1 Csf1r Ctsh Cyp2c44 Cyp2d10 Cyp2d9 Cyp2j6 Cyp39a1 Cyp4v3 D10Ertd641e D16H22S680E D1Ertd622e D2Ertd391e  8 15 2 4 20 10 71 19 11 23 24 1 3 22 128 8 25 3 35 25 28 3 2 0 3 19 7 0 5 2 3 61 3 8 7  11 9 33 8 93 0 200 32 14 12 13 1 8 12 121 14 61 2 103 115 84 68 3 5 35 175 80 0 4 4 33 50 14 25 5  9 15 86 6 81 3 343 31 28 5 49 5 13 20 127 16 120 4 80 106 107 176 3 9 88 355 136 12 2 6 47 45 5 10 9  7 12 99 10 27 9 133 63 21 20 18 11 28 24 98 14 68 6 56 89 88 114 4 18 69 228 141 19 8 13 25 28 6 8 11  37 40 150 30 175 15 1145 158 94 38 143 13 123 60 286 76 388 13 200 600 626 357 11 19 212 1638 265 57 23 14 142 154 22 43 15  239  D630039A03Rik D730039F16Rik Daglb Decr1 Dera Dhdh Dhrs3 Dhrs7 Dlst Dmd Dmgdh Dnase2b Dtx3l Dvl1 Ebp Ebpl Ece1 Edem2 EG240549 Egln3 Ehbp1 Ehd3 Elk3 Elmo3 Elovl5 Enho Enpep Epas1 Eps8l2 Erap1 Ethe1 Etnk2 F13b F2 Fabp1  1 4 2 6 22 4 14 10 101 4 8 0 0 6 7 34 55 6 2 4 6 3 6 2 28 17 12 8 2 5 14 7 2 216 47  0 4 2 0 6 4 16 36 116 11 15 0 8 6 8 44 91 11 0 0 3 21 3 5 2 7 2 11 3 17 33 12 12 508 328  3 16 4 3 6 5 17 47 87 10 18 5 9 13 41 76 92 7 10 4 1 38 0 8 5 13 3 23 5 13 43 21 23 255 421  0 19 4 11 14 6 44 35 46 15 33 6 14 9 19 25 53 5 43 7 7 47 8 7 16 6 13 55 11 10 28 16 61 704 512  11 25 14 16 61 32 70 200 322 24 114 15 32 27 83 458 237 28 87 13 11 88 15 17 62 35 25 130 19 55 324 66 203 872 2156  Fahd1 Fam105b Fam107b Fam131c Fam134a Fam55b Fas Fbxo3 Foxq1 Fpgs Gas2 Gbe1 Gcc1 Gclm Gfm1 Ghr Glo1 Gltpd2 Glul Gm608 Gnmt Gns Got1 Gpld1 Gpsm3 Gpx1 Gramd1c Gstm1 Gstm6 Gstt3 H2-Q10 H6pd Haao Hal Hbp1  25 14 10 2 27 0 1 55 0 24 3 10 8 117 29 18 45 8 294 8 17 59 28 3 4 81 2 355 10 7 34 5 3 3 39  23 15 2 1 31 4 4 100 3 30 7 0 1 21 15 11 6 8 283 5 11 51 30 68 5 8 11 969 15 11 168 4 15 16 20  18 26 4 0 45 2 7 71 2 31 2 2 2 70 14 9 21 17 660 3 65 32 28 113 2 9 9 1027 21 16 227 8 146 29 40  22 7 16 0 41 2 9 57 19 14 1 19 10 132 7 33 41 18 372 11 99 38 49 142 6 58 7 310 28 11 267 15 152 10 64  163 89 38 10 314 12 22 205 22 249 38 19 18 180 68 116 112 28 1212 24 173 189 217 251 16 156 90 1971 81 66 974 21 259 184 136  240  Hc Hdac11 Herpud1 Hfe2 Hiat1 Hmgcl Hmgcs2 Hnf1b Hpd Hpgd Hpx Ifih1 Igfals Igfbp4 Igsf5 Igtp Il13ra1 Il18 Il6st Immp2l Inhbc Iqgap2 Irf7 Irgm2 Isoc2b Itfg1 Itih4 Itpk1 Klc4 Klf15 Kynu Lamp2 Larp2 Lcat Ldlr  4 6 49 0 67 18 75 8 3 1 7 0 7 419 3 0 0 0 6 4 2 20 1 6 2 38 4 5 8 18 0 60 9 128 45  37 3 79 6 53 8 151 5 5 4 31 5 17 219 21 6 2 5 15 6 4 84 11 16 2 16 32 10 1 13 0 146 9 314 82  85 8 84 16 31 22 286 10 48 4 53 1 20 146 27 13 0 6 13 3 18 120 11 71 4 22 193 15 9 6 4 170 8 370 42  83 6 134 47 39 33 327 3 95 21 53 4 8 233 16 8 15 15 26 3 7 43 9 37 1 43 402 3 8 22 19 319 10 178 37  212 35 167 72 200 38 527 18 235 64 93 13 66 903 65 21 17 55 73 22 29 389 33 209 33 63 1229 24 92 98 51 473 21 944 177  Lgals8 Lmbr1l LOC100046168 LOC100047670 LOC100048021 Lrp1 Lrrc28 Ly6e Madd Man2a1 Map1lc3a Mbd2 Mbnl1 Mdfic Mfn1 Mid1ip1 Mmd Mobkl2b Mocos Mpp1 Msra Myd88 Nat15 Ndufb6 Nfe2l2 Nit2 Nqo1 Nt5e Oat Os9 Otc P2rx4 Paqr9 Parp9 Pcmtd1  5 1 70 14 32 34 4 25 3 37 9 62 33 4 10 9 13 5 1 7 8 4 78 237 6 62 1 16 30 13 36 9 183 0 4  16 0 38 1 10 25 3 22 5 37 10 46 57 8 25 31 7 4 2 22 13 7 107 119 3 67 2 8 7 30 171 12 120 3 15  21 0 46 4 7 29 1 46 5 30 10 42 48 4 6 12 7 5 4 49 5 10 105 124 2 104 2 9 3 27 280 16 106 2 17  14 0 49 7 18 42 3 62 3 81 10 53 118 9 6 29 8 4 10 44 10 25 52 78 9 72 2 21 18 26 303 12 116 6 22  127 10 631 23 64 82 10 181 14 99 69 126 154 14 78 91 65 20 24 82 89 29 499 461 12 246 10 41 170 103 599 28 526 15 84  241  Pctp Pcx Pecr Pemt Perp Pgrmc1 Phf20l1 Phlda1 Phyh Pip4k2c Pipox Pitpnb Pls3 Pnpla7 Pnpla8 Pomt2 Ppap2b Ppl Ppm1k Ppp1r3b Ppp1r3c Pqlc1 Prdx5 Prkd2 Prodh Prosc Prr8 Prss36 Pter Ptms Pxmp2 R3hdm1 Rab32 Rassf5 Rb1cc1  8 6 35 6 3 195 6 78 56 8 11 91 31 10 12 11 65 2 3 0 1 4 29 3 2 25 25 3 17 16 21 24 3 2 7  5 3 46 4 34 79 8 82 336 2 10 96 7 15 13 8 19 6 9 0 0 6 19 3 2 12 31 3 15 44 40 15 4 4 8  18 12 40 7 31 105 12 33 268 3 12 116 7 22 21 10 33 2 15 2 2 9 44 2 0 9 20 4 23 60 88 12 17 4 16  10 21 36 37 25 196 17 191 235 6 38 64 39 25 25 6 124 2 23 16 36 9 43 3 0 13 22 1 36 22 49 12 38 3 16  46 56 141 60 79 389 21 606 2458 14 87 237 271 106 54 38 212 15 119 28 81 80 69 14 18 42 69 13 67 138 458 80 48 10 46  Rbm35b Rdh11 Reps2 Retsat Rexo2 Rft1 Rfx5 Rgn Rgp1 Rn18s Rnf114 Rnf125 Rnf13 Rnf14 Rnft1 Rorc Rtp4 S100a1 S1pr1 S1pr5 Sbk1 Sc5d Scn1b Scnn1a Scrn3 Sdc1 Sdc2 Sec14l2 Sec14l4 Sephs2 Serf2 Serpina1a Serpina1d Serping1 Sgk2  0 7 0 3 17 11 3 26 14 5 148 3 19 37 27 1 1 2 24 0 11 43 2 3 7 167 6 7 2 442 8 4 2 70 3  2 15 1 7 6 9 3 141 23 9 47 0 15 57 7 7 7 5 4 2 4 86 4 0 4 112 11 10 2 576 4 3 5 178 7  6 18 2 4 7 11 0 336 12 5 41 2 23 56 8 8 14 4 10 1 6 62 5 1 3 104 15 8 6 329 6 5 3 133 37  2 7 2 5 4 5 1 392 10 7 74 4 46 46 26 12 27 9 20 0 3 126 5 2 4 63 21 5 9 871 8 14 6 137 23  16 97 11 41 55 32 10 1044 90 20 255 12 63 120 38 15 70 43 35 13 23 326 13 13 18 363 124 20 32 1629 15 18 15 901 62  242  Sgms2 Sgpl1 Sh3d19 Shpk Siae Sigmar1 Skap2 Slc12a9 Slc23a1 Slc25a1 Slc25a16 Slc25a42 Slc26a1 Slc27a2 Slc29a1 Slc30a10 Slc35d2 Slc38a4 Slc46a3 Slc47a1 Slc48a1 Slc6a6 Slc7a10 Smpd1 Smpdl3a Sod1 Sos2 Spast Sqstm1 Srd5a1 Srgap3 Srr Ssfa2 St3gal1 St3gal4  10 39 51 1 5 88 1 8 1 19 17 5 0 20 8 0 0 145 1 1 11 32 18 10 4 861 22 6 63 4 2 3 9 10 9  12 30 21 4 3 66 1 2 5 68 15 9 3 32 9 7 0 193 0 12 28 14 4 7 9 1142 8 2 173 1 4 3 8 8 14  8 33 21 4 8 116 4 4 7 71 18 10 9 29 8 2 6 211 4 22 20 22 6 12 6 1087 10 2 440 1 6 8 3 13 10  9 58 17 5 9 39 5 2 5 11 12 12 4 83 19 7 6 135 6 27 27 23 5 10 15 387 23 3 171 1 0 8 16 23 30  75 119 116 33 47 271 10 15 13 134 130 24 13 166 29 11 16 1667 28 35 64 152 26 21 24 7313 35 16 1063 23 13 15 17 28 88  St6gal1 Stard5 Stom Stx17 Suclg2 Suhw4 Sult1b1 Suox Syngr2 Tars Tek Terf1 Tfb2m Tgoln1 Them2 Thoc2 Tmbim1 Tmbim6 Tmed5 Tmem106a Tmem106b Tmem126b Tmem129 Tmem19 Tmem205 Tmem30b Tmem38b Tmem9b Tmprss6 Tnfaip8l1 Tob1 Trfr2 Trib1 Trp53i11 Trp53inp2  2 11 2 8 2 12 3 13 47 51 2 9 18 82 57 15 10 102 16 0 11 2 3 9 28 0 9 25 19 4 42 9 8 5 4  25 22 2 1 2 11 4 13 23 38 1 9 8 60 40 4 10 42 15 4 30 3 6 19 73 3 4 20 19 11 32 78 2 2 2  13 27 3 5 0 11 1 5 37 22 0 5 6 56 51 5 15 73 14 9 22 6 4 10 187 19 8 26 37 10 54 139 0 3 7  10 47 8 7 1 13 2 4 27 13 7 7 6 115 33 24 8 131 23 11 27 3 2 10 91 34 21 29 45 5 83 127 16 9 11  42 63 14 12 11 27 30 83 94 93 13 31 33 155 146 36 111 204 62 18 100 14 12 100 344 57 26 98 84 21 143 379 16 24 15  243  Tyk2 Ube3b Ugp2 Ulk2 Unc93b1 Upb1 Uroc1 Ushbp1 Vps4a Vtn Wdfy1 Wdr23 Xylb Zbtb4 Zbtb7b Zfp672 Znfx1  1 28 72 17 1 10 2 0 8 224 3 33 13 6 3 10 4  4 27 99 17 2 10 8 0 12 350 3 7 7 7 8 6 2  3 31 137 25 5 19 13 1 8 521 5 7 3 5 13 4 4  1 14 158 23 6 38 80 6 4 404 6 20 3 23 4 4 7  12 67 767 66 26 47 102 12 26 2318 17 102 39 27 27 19 20  244  Appendix N Cluster N genes and their expression in the five Tag-seq libraries (in tags per million).  Symbol 2210023G05Rik 2810007J24Rik 5033411D12Rik 5730409E04Rik 9030617O03Rik Abca8a Abcb11 Abcb4 Abcc3 Acaa1b Acat3 Acsm5 Adh1 Agtr1a AI132487 AI317395 Akr1c14 Akr1d1 Aldh1a7 Aldh9a1 Ang Angptl3 Aox1 Aox3 Apoc4 Apol7a Aqp1 Aqp9 Atp6v0d2  E10 DLK1+ 0 1 0 0 0 0 0 0 0 0 0 0 0 3 4 1 0 0 0 16 1 8 0 0 5 0 0 0 1  E12 DLK1+ 0 0 2 0 0 0 2 2 0 17 2 0 0 7 11 0 0 2 0 12 3 140 0 0 19 0 0 5 0  E14 DLK1+ 0 1 6 2 0 0 10 4 0 103 10 0 4 10 27 1 0 4 1 11 13 144 0 0 96 0 2 3 1  E16 DLK1+ 5 3 12 0 2 1 31 5 8 97 23 0 25 27 51 1 14 61 1 12 8 157 1 7 347 2 7 3 1  adult liver 40 69 61 14 15 31 138 69 25 2808 88 10 280 134 1476 18 401 182 32 381 360 1597 12 72 1661 27 44 165 11  Azgp1 Baat BC026782 BC029214 Bco2 C4bp C8g Camk1d Cdo1 Ces1 Ces3 Chpt1 Cldn2 Clec2d Clec4f Cmbl Col15a1 Ctso Cyp17a1 Cyp27a1 Cyp2a22 Cyp2r1 Cyp3a11 Cyp3a13 Cyp7a1 Cyp8b1 Dnase1l3 Dpyd Egfr Ehhadh Ephx1 F9 Fam134b Fbp1 Fgl1  0 0 0 4 0 0 10 2 53 0 0 3 0 1 0 1 0 1 0 0 0 1 0 2 0 0 0 1 0 7 0 0 0 0 1  0 0 2 0 0 4 24 6 68 2 3 2 0 12 2 6 0 0 0 0 0 0 0 1 0 0 0 3 0 13 17 0 0 0 0  3 5 15 6 0 2 38 5 77 3 7 3 0 16 7 13 0 2 0 1 1 1 6 5 1 3 2 0 0 22 36 2 0 1 0  93 11 95 6 5 9 43 17 59 7 22 3 1 4 34 6 3 11 3 6 6 1 44 8 1 5 5 2 1 40 19 21 6 52 7  743 184 527 127 19 189 889 90 1007 259 387 77 11 148 104 169 14 80 37 119 91 18 1244 99 22 84 69 66 21 228 1015 78 21 175 192  245  Fmo1 Gabarapl1 Gch1 Gck Gjb2 Gpc4 Gpihbp1 Gpr116 Gpt Gstk1 H2-Aa H2-Ab1 Hao1 Hp Hspb8 Ifi27l2a Ifit1 Igfbp2 Igsf11 Iigp1 Il1rap Inhbe Insig2 Itpr1 Kmo Mat1a Mbl2 Mgll Mgst1 Mrap Myo1b Nr0b2 Papss2 Parp14 Pck1  0 12 6 0 2 0 0 1 4 2 0 1 0 1 1 0 0 11 1 0 0 0 5 12 1 0 0 0 17 1 19 3 2 0 0  0 9 21 0 4 0 0 0 0 2 0 0 0 3 2 0 0 0 0 0 0 0 6 5 18 3 0 0 623 0 13 2 0 0 0  3 18 17 0 40 0 0 2 4 3 0 3 0 28 6 1 2 3 0 0 0 0 5 6 29 42 1 1 1379 0 16 0 3 0 0  15 27 12 1 72 7 5 7 4 5 1 2 10 52 13 5 2 2 0 2 1 0 3 14 29 120 4 12 729 3 19 0 16 4 8  87 284 327 10 590 50 37 32 79 105 33 34 33 471 129 47 18 379 18 48 19 10 143 139 271 1273 84 99 12376 32 238 42 67 11 170  Plscr2 Pzp Rarres2 Rdh7 Rnase4 Rnd1 Rxrg Scd1 Sdr42e1 Sepx1 Slc17a2 Slc27a5 Slc38a3 Slco2a1 Spnb3 Spp1 Steap4 Sult1d1 Tmem184a Tmem56 Ttc39b Ube2u Ugt2b36 Ugt2b5 Uox Vipr1 Vnn3 Zhx2 1100001G20Rik 2200001I15Rik 9030619P08Rik 9530008L14Rik A1bg Adh4 Agxt2l1  0 13 9 0 12 0 3 30 1 0 0 1 9 1 1 1 0 1 1 1 0 1 1 1 0 0 0 2 0 0 0 0 0 0 0  0 234 5 0 72 0 0 55 11 0 2 47 36 2 0 2 0 6 0 0 0 0 11 3 0 0 0 0 0 0 0 0 0 0 0  0 521 16 17 65 0 0 34 7 1 2 106 90 12 0 0 3 2 0 0 1 0 39 7 7 0 0 0 0 0 0 0 0 1 0  0 1082 96 149 44 4 0 85 11 1 6 44 111 6 0 2 21 10 0 11 1 0 127 15 70 1 1 3 2 0 0 0 0 0 0  11 4881 697 2613 741 22 28 2137 170 11 116 1092 878 83 16 18 82 66 26 56 14 27 438 126 402 12 10 22 764 32 146 18 2114 192 29  246  Akr1c6 Apoa4 Apoa5 Apoc3 Apol9b Apon Asgr1 AU018778 BC014805 BC089597 Bdh2 C730048C13Rik Cml2 Cyp1a2 Cyp2a12 Cyp2c50 Cyp2c67 Cyp2c70 Cyp2e1 Cyp2f2 Cyp2j5 Cyp4a14 Cyp4f14 Dio1 EG13909 Es1 Fmo3 G6pc Glyat Hacl1 Hnmt Hrg Inmt Itga7 Lrg1  0 0 2 10 0 0 3 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 2 0 0  0 0 0 43 0 0 6 0 0 0 0 0 0 0 6 0 3 0 0 0 0 0 0 0 0 0 0 0 0 8 0 0 0 0 0  3 0 0 23 0 0 9 0 0 0 0 0 0 2 5 0 2 1 0 0 0 0 0 0 0 1 0 0 0 22 0 0 1 0 0  6 0 2 20 0 2 7 0 0 1 0 0 0 1 16 0 1 0 0 2 1 0 0 0 0 12 0 1 1 13 0 2 0 0 1  681 81 615 2488 48 284 1184 32 152 519 20 69 22 880 885 510 354 34 12 872 239 59 67 40 1161 1251 246 129 94 1138 16 189 1003 13 110  Ly6a Mettl7b Mme Mup3 Ndrg2 Nox4 Odf3b Orm3 OTTMUSG00000007431 OTTMUSG00000007480 Pon1 Prlr Rdh16 Rgl3 Sds Serpina3c Serpina3k Serpina3m Slc22a1 Slc34a2 Slco1a1 Slco1b2 Slco2b1 Sult2a1 Tat Tdo2 Thrsp Zap70  0 0 0 0 56 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0  0 0 0 0 232 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0  0 0 0 0 200 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 6 0 0 0 0 0 0  1 0 0 0 150 0 0 0 0 0 7 0 0 0 0 0 0 0 0 0 0 33 0 6 0 0 0 0  38 42 13 2642 22427 15 11 12 604 947 1341 62 100 14 54 123 9792 170 417 11 28 2321 80 215 287 678 167 15  247  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0073493/manifest

Comment

Related Items