Open Collections

UBC Faculty Research and Publications

Genome-wide survey reveals dynamic widespread tissue-specific changes in DNA methylation during development Liang, Ping; Song, Fei; Ghosh, Srimoyee; Morien, Evan; Qin, Maochun; Mahmood, Saleh; Fujiwara, Kyoko; Igarashi, Jun; Nagase, Hiroki; Held, William A May 11, 2011

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

Item Metadata

Download

Media
52383-12864_2010_Article_3387.pdf [ 1.19MB ]
Metadata
JSON: 52383-1.0223776.json
JSON-LD: 52383-1.0223776-ld.json
RDF/XML (Pretty): 52383-1.0223776-rdf.xml
RDF/JSON: 52383-1.0223776-rdf.json
Turtle: 52383-1.0223776-turtle.txt
N-Triples: 52383-1.0223776-rdf-ntriples.txt
Original Record: 52383-1.0223776-source.json
Full Text
52383-1.0223776-fulltext.txt
Citation
52383-1.0223776.ris

Full Text

RESEARCH ARTICLE Open AccessGenome-wide survey reveals dynamicwidespread tissue-specific changes in DNAmethylation during developmentPing Liang1*†, Fei Song2,3†, Srimoyee Ghosh4,5, Evan Morien4,6, Maochun Qin4, Saleh Mahmood2,7,Kyoko Fujiwara4,8, Jun Igarashi8, Hiroki Nagase4,8,9* and William A Held2AbstractBackground: Changes in DNA methylation in the mammalian genome during development are frequent events andplay major roles regulating gene expression and other developmental processes. It is necessary to identify these eventsso that we may understand how these changes affect normal development and how aberrant changes may impactdisease.Results: In this study Methylated DNA ImmunoPrecipitation (MeDIP) was used in conjunction with a NimbleGenpromoter plus CpG island (CpGi) array to identify Tissue and Developmental Stage specific Differentially MethylatedDNA Regions (T-DMRs and DS-DMRs) on a genome-wide basis. Four tissues (brain, heart, liver, and testis) from C57BL/6Jmice were analyzed at three developmental stages (15 day embryo, E15; new born, NB; 12 week adult, AD). Almost5,000 adult T-DMRs and 10,000 DS-DMRs were identified. Surprisingly, almost all DS-DMRs were tissue specific (i.e.methylated in at least one tissue and unmethylated in one or more tissues). In addition our results indicate that manyDS-DMRs are methylated at early development stages (E15 and NB) but are unmethylated in adult. There is a verystrong bias for testis specific methylation in non-CpGi promoter regions (94%). Although the majority of T-DMRs andDS-DMRs tended to be in non-CpGi promoter regions, a relatively large number were also located in CpGi in promoter,intragenic and intergenic regions (>15% of the 15,979 CpGi on the array).Conclusions: Our data suggests the vast majority of unique sequence DNA methylation has tissue specificity, thatdemethylation has a prominent role in tissue differentiation, and that DNA methylation has regulatory roles inalternative promoter selection and in non-promoter regions. Overall, our studies indicate changes in DNAmethylation during development are a dynamic, widespread, and tissue-specific process involving both DNAmethylation and demethylation.BackgroundDNA methylation in mammals occurs predominantly atCpG dinucleotides [1]. Early studies indicated that innormal cells, most CpGs in repeats, retroviral sequences,and within the coding region of genes are methylated,whereas most CpG island (CpGi) regions are maintainedin an unmethylated state [2,3]. After fertilization there isa period of genome-wide demethylation followed by per-iods of remethylation in somatic cells after implantation.Demethylation and reprogramming is also necessary inthe germ cells to reset gamete-specific imprinting [4-7].There is a long history and generally inverse correlationbetween gene promoter DNA methylation and geneexpression [8-12]. Early studies proposed that regulatedmethylation and demethylation has a role regulatinggene expression during development [10,13].The critical function of DNA methylation during devel-opment is also apparent from the consequence of targetedknockouts of DNA methyltransferase genes. The knockoutof Dnmt1, which has a strong preference for hemimethy-lated DNA and is thus considered to be a maintenance* Correspondence: pliang@brocku.ca; hnagase@chiba-cc.jp† Contributed equally1Department of Biological Sciences, Brock University, St. Catharines, Ontario,Canada4Department of Cancer Genetics, Roswell Park Cancer Institute, Buffalo, NY,USAFull list of author information is available at the end of the articleLiang et al. BMC Genomics 2011, 12:231http://www.biomedcentral.com/1471-2164/12/231© 2011 Liang et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative CommonsAttribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction inany medium, provided the original work is properly cited.DNA methyltransferase, is lethal during development [14].Dnmt3a and Dnmt3b are required for de novo methylationduring development [15]. The persistent expression ofboth Dnmt 3a and Dnmt 3b after gastrulation and implan-tation suggests that de novo methylation occurs duringlater stages of development.Restriction Landmark Genomic Scanning (RLGS)detected differences in DNA methylation between tissuesmore than 15 years ago [16-19]. These studies, as well asmore recent reports using a variety of methods, havemarkedly altered our concept of differentially methylatedregions (DMRs) in the mammalian genome and the pro-gression of DNA methylation changes during development[19-36]. Tissue Specific Differentially Methylated Regions(T-DMRs) include both CpGi and non-CpGi promoterregions as well as intragenic and intergenic regions[19,32,37,38]. Analysis of T-DMRs using a CpGi arrayindicated that T-DMRs in CpGi regions are associatedwith developmental gene loci and many are located innon-promoter regions [23]. CpGi amplification in con-junction with microarrays, indicated approximately 4% ofdense CpGi promoters are methylated in normal periph-eral blood [31]. A number of studies have consistentlyobserved that weak or non-CpGi promoters are major tar-gets for tissue specific DNA methylation [12,28,30,33,39].Analysis of DNA methylation changes during develop-ment in vitro and in vivo are more limited but indicatethat DNA methylation is a dynamic process involvingboth methylation and demethylation [20,26,32]. In vivoanalysis of methylation changes within a tissue at differ-ent developmental stages is complicated by changes incell populations [32]. In vitro analysis of DNA methyla-tion during differentiation of ES cells or progenitor cellsis complicated by potential aberrant methylation due togrowth of cells in tissue culture [20,26,40].Our previous results identified a limited number ofT-DMRs using RLGS, suggesting that methylationchanges during development are dynamic and involveboth methylation and demethylation [19,32]. The resultsreported here investigate this with a more comprehen-sive set of T-DMRs and Developmental Stage specificDifferentially Methylated DNA Regions (DS-DMRs)identified through Methylated DNA ImmunoPrecipita-tion (MeDIP) [20,32].ResultsMeDIP methylation analysisThe MeDIP methylation analysis developed by Weber et al[33] was utilized to identify genomic sites of differentialDNA methylation in selected 12 wk adult C57BL/6Jmouse tissues (T-DMRs) and in tissues at different devel-opmental stages (DS-DMRs: 15 day embryo, E15; newborn, NB; and 12 wk adult, AD). Methylated DNA wasimmunoprecipitated with antibody to 5-methyl-cytidine,the input and immunoprecipitated (IP) DNA were differ-entially labeled and hybridized to a NimbleGen promoterplus CpGi array according to a protocol provided by Nim-bleGen. The NimbleGen Array covers all UCSC-annotatedCpGi (15,979) as well as promoters for all RefSeq genes(19,530). Repetitive regions are excluded from the Array.Non-CpGi intragenic and intergenic regions are not cov-ered by the array except in some limited regions in whichthe array is essentially a tiling array (1.9 Mb on portions ofchromosomes 6, 7, and 17).Assessment of the MeDIP data quality and validation ofMeDIP methylation peaksA number of methods were used to validate the quality ofthe MeDIP methylation data. These include scatter plotsof log 2 ratios between IP DNA and input DNA, analysisof Pearson Coefficients between biological replicate sam-ples, comparison of MeDIP methylation with previouslyidentified RLGS T-DMRs, MeDIP analysis of other knownmethylation sites, such as imprinted genes, and analysis ofboth selected and randomly chosen MeDIP methylatedregions by Sequenom MassARRAY quantitative methyla-tion analysis of bisulfite treated DNA. An example of aselected region analyzed by Sequenom MassARRAY isshown in Figure 1. The results indicate a close similaritybetween MeDIP and Sequenom MassARRAY methylationanalysis. Additional details of Sequenom MassARRAYmethylation analysis and other validation methods are pre-sented in the Methods section and Additional files.Identification of genomic sites of DNA methylation inadult tissues (brain, heart, liver, and testis)The IP and input DNA from 12 wk adult C57BL/6J mousetissues (brain, heart, liver, and testis) were hybridized to aNimbleGen Promoter plus CpGi array. T-DMRs are geno-mic regions that are methylated in one or more of the fouradult tissues, but unmethylated in at least one tissue. Forthe identification of T-DMRs, the log2 ratios of all probesfor each sample were normalized and used to identifymethylation peaks as described in the methods. To facili-tate the analysis of MeDIP data, an ad hoc bioinformaticstool was developed to process the methylation peak listsfor individual samples and identify genomic regions thatare 1) commonly methylated in all tissues; 2) methylatedin a single tissue; and 3) methylated in some but not alltissues. Only those methylation peak regions that werepresent in both biological replicate samples and wereabsent from at least one other tissue were designated asT-DMRs. Thus, the frequency of T-DMRs may be under-estimated. T-DMR locations were designated as 1) CpGiPromoter; 2) Intragenic CpGi; 3) Intergenic CpGi; and 4)Non-CpGi promoter (see details in Methods).Summaries of the number and distributions of theadult T-DMRs are shown in Table 1. Overall, there areLiang et al. BMC Genomics 2011, 12:231http://www.biomedcentral.com/1471-2164/12/231Page 2 of 16almost 5,000 adult T-DMRs that are present in one ormore of the examined adult tissues but are absent inone or more tissues, whereas there are only 460 methy-lated regions that are common to all four tissues. Thereare a relatively high number of T-DMRs unique to testis(1,294) and a relatively low number unique to brain(285). Most of the T-DMRs (69.3%) are located withinnon-CpGi promoter regions, but 30.7% are within CpGiregions, among which 13.7% are within CpGi promoterregions and 17% are in non-promoter CpGi regions(intragenic and intergenic CpGi regions). Within the dif-ferent tissues there are some notable differences. ForCpG islandTranscriptTranscript StartLiver 1Liver 2Testis 2Testis 1ES 1ES 2Scaled log2 ratioStk31Sequenom Methylation AnalysisMethylated TestisLiver 2Testis 2ES 1ES 2Testis 10 Liver 1100% Not analyzed Figure 1 Comparison of methylation data between MeDIP/NimbleGen Promoter + CpGi Array and Sequenom MassArray: Stk31 CpGipromoter region. The Stk31 gene region with a CpGi promoter region that is methylated in the liver, but not in ES or testis cells. The boxindicates the position of the methylation peak in the liver that overlaps with the CpGi promoter region analyzed by Sequenom. In the epigramof Sequenom methylation analysis shown at the bottom panel, the blue circles indicate 100% methylation, green circles indicate around 50%methylation and yellow circles indicate 0% methylation. The results from biological duplicate samples are shown. The bottom lane (SequenomMethylation Analysis) shows the results of analysis of testis DNA after in vitro methylation using M. Sss1 methylase (New England Biolabs).Liang et al. BMC Genomics 2011, 12:231http://www.biomedcentral.com/1471-2164/12/231Page 3 of 16example, approximately 94% of the testis-specificT-DMRs are in non-CpGi promoter regions, whereaseach of the three somatic tissues is no more than 60%.T-DMRs that are present in brain, heart, and liver, butnot testis, would include somatic methylation sites. Themajority of these (B+H+L) are in CpGi regions (64.3%).The number and distribution of T-DMRs in ES cellsES cells are derived from the inner cell mass of the earlyblastocyst that gives rise to the embryo. Following fertiliza-tion, the embryonic cells undergo a period of demethyla-tion and make up/form the inner cell mass. Remethylationof somatic tissues is thought to resume following implan-tation [7]. Thus, we compared the methylation status ofT-DMRs in E15 tissues with ES cells (Additional file 1).Overall, there are fewer T-DMRs (2750) from the compar-ison of ES cell and E15 tissues than T-DMRs in adult tis-sues (4686, Table 1). We found more unique methylationpeaks in ES cells (981) than in E15 tissues of brain (310),heart (131), liver (764), or testis (230). This indicates thatthere are many sites that are methylated in ES cells but areunmethylated in E15 tissues with most (85.5%) located innon-CpGi promoter regions.T-DMRs in regions outside CpGi regions and promotersThe NimbleGen promoter + CpGi array contains a tilingarray that covers about 1.9 Mb of the mouse genome. Thisincludes the Hoxa gene cluster on chromosome 6 (423 Kb),the Igf2r imprinted region on chromosome 17 (88 Kb) andimprinted regions on chromosome 7 (1,390 Kb). In theseregions it is possible to identify tissue specific methylationpeaks that are not associated with CpGi regions or promo-ters. For example, in the KvDMR region on chromosome 7(Additional file 2), there are non-CpGi and non-promotermethylation peaks in testis and brain that are not present inliver in addition to T-DMRs in CpGi regions. The resultsshown in Additional file 3a indicate that there is similar dis-tribution of T-DMRs in the tiling regions in CpGi and non-CpGi promoter regions as in the rest of the array-coveredgenome. However, there are additional methylation peakswithin the tiling regions (Additional file 3b) in non-CpGiintragenic regions (15.7%) and in non-CpGi intergenicregions (12.9%). These results indicate that almost 30% ofthe T-DMRs may be missed by restricting analysis to CpGiand promoter regions.Developmental Stage-Specific Differentially MethylatedRegions (DS-DMRs)We used MeDIP to determine the number and location ofmethylated sites at the different developmental stageswithin each tissue (Table 2). DS-DMRs are genomic regionsthat are methylated at one or more developmental stage ofa tissue, but unmethylated in at least one stage. It should benoted that DS-DMRs are selected as stage-specific differen-tially methylated regions within a tissue without regard topresence or absence in another tissue. This analysis pro-vides a basis for determining changes in developmentalstage specific methylation that occur within a tissue. Thesechanges may reflect active or passive changes in DNAmethylation as well as changes in the proportion of specificcell populations. The Pearson Coefficients between theduplicate biological replicates varied somewhat fromTable 1 Adult T-DMRsTissue Total CpGi Prom % Non-CpGi Prom % Intra-genic CpGi % Inter-genic CpGi %Common 460 18 3.9 230 50.0 163 35.4 49 10.7Brain unique 285 68 23.9 173 60.7 23 8.1 21 7.4Heart unique 822 223 27.1 428 52.1 115 14.0 56 6.8Liver unique 711 155 21.8 428 60.2 66 9.3 62 8.7Testis unique 1294 44 3.4 1215 93.9 20 1.5 15 1.2Total T-unique 3112 490 15.7 2244 72.1 224 7.2 154 4.9B + H 482 69 14.3 282 58.5 89 18.5 42 8.7B + L 12 2 16.7 7 58.3 2 16.7 1 8.3B + T 26 1 3.8 25 96.2 0 0.0 0 0.0H + L 49 7 14.3 24 49.0 13 26.5 5 10.2H + T 47 0 0.0 45 95.7 2 4.3 0 0.0L + T 305 3 1.0 268 87.9 27 8.9 7 2.3B + H + L 414 65 15.7 148 35.7 127 30.7 74 17.9B + H + T 175 3 1.7 151 86.3 19 10.9 2 1.1B + L + T 18 0 0.0 15 83.3 2 11.1 1 5.6H + L + T 46 4 8.7 37 80.4 4 8.7 1 2.2Totalmulti-tissues 1574 154 9.8 1002 63.7 285 18.1 133 8.4Total T-DMRs 4686 644 13.7 3246 69.3 509 10.9 287 6.1Liang et al. BMC Genomics 2011, 12:231http://www.biomedcentral.com/1471-2164/12/231Page 4 of 16Table 2 The number of DS-DMRs in different tissuesTotal Adjusted# %1 CpGi-Prom %2 non-CpGi-Prom %2 Intra-genic CpGi %2 Inter-genic CpGi %2Brain:Common 1903 NA ND 182 9.6 1018 53.5 493 25.9 210 11.0E15_uniq 533 533 27.5 105 19.7 359 67.4 35 6.6 34 6.4E15+NB 490 509 25.3 66 13.5 368 75.1 29 5.9 27 5.5NB_uniq 370 400 19.1 55 14.9 248 67.0 42 11.4 25 6.8AD_uniq 408 532 21.1 94 23.0 233 57.1 52 12.7 29 7.1NB+AD 70 82 3.6 5 7.1 40 57.1 16 22.9 9 12.9E15+AD 65 77 3.4 11 16.9 46 70.8 7 10.8 1 1.5Total* 1936 2132 100.0 336 15.2 1294 62.6 181 13.9 125 8.4UnMe-AD** 1393 1442 72.0 226 16.2 975 70.0 106 7.6 86 6.2Me-AD*** 543 691 28.0 110 20.3 319 58.8 75 13.8 39 7.2Heart:Common 1844 NA ND 218 11.8 946 51.3 491 26.6 189 10.2E15_uniq 46 81 6.9 27 58.7 8 17.4 2 4.3 9 19.6E15+NB 87 125 13.0 17 19.5 61 70.1 4 4.6 5 5.7NB_uniq 93 112 13.9 12 12.9 64 68.8 7 7.5 10 10.8AD_uniq 172 172 25.6 45 26.2 83 48.3 33 19.2 11 6.4NB+AD 246 253 36.7 44 17.9 132 53.7 54 22.0 16 6.5E15+AD 27 34 4.0 6 22.2 15 55.6 5 18.5 1 3.7Total* 671 777 100.0 151 22.4 363 54.1 105 15.6 52 7.7UnMe-AD** 226 318 33.7 56 24.8 133 58.8 13 5.8 24 10.6Me-AD*** 445 459 66.3 95 21.3 230 51.7 92 20.7 28 6.3Liver:Common 1497 NA ND 141 9.4 726 48.5 425 28.4 205 13.7E15_uniq 117 119 4.9 25 21.4 74 63.2 10 8.5 8 6.8E15+NB 1419 1431 60.0 198 14.0 1034 72.9 125 8.8 62 4.4NB_uniq 147 147 6.2 24 16.3 109 74.1 8 5.4 6 4.1AD_uniq 468 813 19.7 70 15.0 322 68.8 45 9.6 31 6.6NB+AD 36 46 1.5 2 5.6 30 83.3 3 8.3 1 2.8E15+AD 194 249 8.1 13 6.7 130 67.0 31 16.0 20 10.3Total* 2381 2804 100.0 332 13.9 1699 71.4 222 9.3 128 5.4UnMe-AD** 1683 1696 70.7 247 14.7 1217 72.3 143 8.5 76 4.5Me AD*** 698 1107 29.3 85 12. 482 69.1 79 11.3 52 7.4Testis:Common 570 NA ND 20 3.5 363 63.7 148 26.0 39 6.8E15_uniq 272 380 6.3 105 38.6 101 37.1 33 12.1 33 12.1E15+NB 939 1094 22.0 193 20.6 480 51.1 157 16.7 109 11.6NB_uniq 798 789 18.4 231 28.9 391 49.0 86 10.8 90 11.3AD_uniq 2214 2336 50.9 71 3.2 1908 86.2 165 7.5 70 3.2NB+AD 63 65 1.4 3 4.8 48 76.2 9 14.3 3 4.8E15+AD 62 91 1.4 5 8.1 38 61.3 14 22.6 5 8.1Total 4348 4755 100.0 608 14.0 2966 68.2 464 10.7 310 7.1UnMe-AD** 2009 2263 40.8 529 26.3 972 48.4 276 13.7 232 11.5Me-AD*** 2339 2493 47.6 79 3.4 1994 85.3 188 8.0 78 3.3Grand Total:Common 5814 NA ND 561 9.6 3053 52.5 1557 26.8 643 11.1DS-DMRs 9336 10469 100.0 1427 15.3 6322 67.7 972 10.4 615 6.6UnMe-AD** 5311 5719 56.9 1058 19.9 3297 62.1 538 10.1 418 7.9Me-AD*** 4025 4750 43.1 369 9.2 3025 75.2 434 10.8 197 4.9%1: percentage is based on the row total vs. total DMRs; %2: percentage is based on the total of DMRs in the row; Total*: Stage specific; UnMe-AD**:unmethylated in AD but methylated in E15 and/or NB; Me-AD***: methylated in Adult but unmethylated in E15 and/or NB; Adjusted#: number of DMRs adjustedbased on the adjusted coefficiency between the duplicated samples.Liang et al. BMC Genomics 2011, 12:231http://www.biomedcentral.com/1471-2164/12/231Page 5 of 16sample to sample (Additional file 4). Therefore the totalnumber of methylation peaks would be somewhat underes-timated in samples with low relative to those with highPearson Coefficients. We adjusted the total number ofmethylation peaks as if the Pearson Coefficients for allstages were the same as the best one in that tissue (see datalabeled “Adjusted” in Table 2). Although the total numberof peaks increases for the samples with lower Coefficients,the overall trends are comparable (Table 2 and Figure 2),indicating that our underestimation of methylation peaksdoes not have a significant impact on the distribution ofmethylation peaks into different genomic regions.There were considerably more methylation sites thatwere common within a tissue at the different develop-mental stages than between different adult tissues (460common sites, Table 1). Brain, heart, and liver each had1,903, 1,844 and 1,497 common methylation sites,respectively, whereas testis had 570 (Table 2). Similar tothe common sites for the adult tissues, about 50% ofdevelopmental stage common sites are in CpGi regions(CpGi promoter + intragenic and intergenic CpGi) andabout 50% are in non-CpGi promoter regions (Table 1and 2). Somewhat surprising is the relatively high num-ber of sites in liver (and brain and testis, to somedegree) that are methylated in E15 + NB but unmethy-lated in adult (1,419). Together, in somatic tissues(heart, brain, and liver), there were almost 5,000 sites ofmethylation that exhibit stage specific differences. Testisalone had 4,348 sites of stage specific methylatedregions, most of which were methylated in adult andlocalized to non-CpGi promoters. Again, as for theadult tissues, the most frequent location for differentialmethylation was within non-CpGi promoter regions(54.1% to 71.4%), followed by CpGi promoter regions(13.9 to 22.4%), intragenic CpGi regions (9.3 to 15.6%),and intergenic CpGi regions (5.4 to 8.4%). In contrast toother tissues, the location of testis stage specific DS-DMRs appear to shift from a high proportion of CpGilocation at early developmental stages (E15, 62.9% andNB, 51.0%) to a high proportion of non-CpGi promoterlocation in adult (86.2%).Many DS-DMRs are unmethylated in adultWe grouped the DS-DMRs within a tissue according towhether they were unmethylated in adult (UnMe-AD)but methylated at earlier developmental stages (E15, E15+NB, NB). These would represent sites of demethylationduring development. Conversely sites that were methy-lated in adult (E15+AD, NB+AD, AD) but unmethylatedat earlier developmental stages would be sites that weremethylated during development. Surprisingly, there weremore sites that became unmethylated (5,311) thanbecame methylated (4,025) during development (Figure 2and Table 2). Adjustment for differences in the PearsonCoefficient (Figure 2) does not significantly alter this gen-eral conclusion. Among the different tissues, there seemto be some significant differences with regard to the ratioof DMRs subjected to methylation and demethylation. Intestis, there are an almost equal number of sites thatbecome unmethylated (2,009) or methylated (2,339) inadult. However, in brain and liver there are a much largernumber of sites undergoing demethylation in adult thanthat of methylation whereas in heart, there are a smallernumber of sites undergoing demethylation than methyla-tion. These results indicate dynamic changes in DNAmethylation during development and that epigeneticchanges resulting in unmethylated DS-DMRs are a majorfeature of differentiation into adult tissues.Overlap between DS-DMRs in different tissuesThe DS-DMRs indicate differential stage-specific sites ofmethylation within a tissue without regard to whether it istissue specific. To determine how many of the DS-DMRsare tissue specific, we determined the extent of overlappingDS-DMR (i.e., methylation) between tissues (Table 3). Theresults indicate for all four tissues together, the majority(7,857 or 80.3%) of the 9,784 DS-DMRs are unique to asingle tissue whereas the remaining 1,927 DS-DMRs arepresent in two or more but not all tissues, and thus arealso tissue-specific. Together, it indicates that all of the9,784 DS-DMRs are also T-DMRs. If testis is excluded,there are 4,232 out of 5,519 DS-DMRs that are unique to asingle tissue and 613 peaks being present from one ormore but not all tissues (Additional file 5a). In this case, wedid find a small number of peaks (21) that are common to50010001500200025003000Number of peaksTotal numberAdjusted total numberUnmethylated MethylatedHeart    Brain      Liver   Testis     Heart   Brain    Liver      TestisFigure 2 DS-DMRs: Number of peaks that are Unmethylated orMethylated in adult Tissues. Total numbers of unmethylatedpeaks and methylated peaks are shown in adult heart, brain, liverand testis. The grey bar indicates the actual total number ofunmethylated and methylated peaks in each tissue. The black barindicates total number of unmethylated and methylated peaks ineach tissue after adjustment for differences in the PearsonCoefficient.Liang et al. BMC Genomics 2011, 12:231http://www.biomedcentral.com/1471-2164/12/231Page 6 of 16all three tissues, suggesting that the 3 somatic tissues sharemore commonality among each other than with testis. Wealso examined the overlap for the 5,761 DMRs that arecommon to all stages in a tissue (Additional file 5b). Themajority of these DMRs are also tissue-specific (5,487/5,761 or 95.2%). However, compared to DS-DMRs, the per-centage of these DMRs being tissue-unique is much lower,approximately 15% for each of the somatic tissues and only1.9% for testis. Only a small proportion (274 or 4.8%) arecommon to all developmental stages and all tissues. Over-all, since all of the 9,784 DS-DMRs and 5,487 of the 5,761DMRs common to all developmental stages in a tissue arealso T-DMRs, we identified a total of 15,271 T-DMRs.Tissue-Specific differences in DNA methylation within theHoxa gene clusterHomeobox genes have important roles in regulating devel-opment [41,42]. One of the tiled regions available on theNimbleGen array includes the Hoxa gene cluster on chro-mosome 6. Additional file 6 shows the scaled log2 ratiosfrom the NimbleGen SignalMap view of an approximately100 Kb region that includes the Hoxa gene cluster. Thereare 8 adult T-DMRs in this region that were present inboth biological replicate DNA samples of the involved tis-sue(s). Even though this region is CpGi rich, none of thepeaks were within CpGi regions and only one was loca-lized in a promoter region (Additional file 7; peak 7,Hoxa7). The other methylation peaks were within introns(peaks 2, 3, and 4) or associated with the 3’ end (exon) ofthe gene (peaks 1, 5, 6, and 8). Four of the peaks (peaks 2-5) were testis specific, one (peak 6) was heart specific, one(peak 7) was present in heart and liver, and one (peak 8)was in heart and brain. All sequences corresponding tothese peaks were highly conserved, suggesting a regulatoryfunction. Analysis of methylation differences within theHoxa gene cluster during development (Figure 3) indicatesthat two additional DS-DMRs not seen as T-DMRs foradult tissues were observed when samples from all threedevelopmental stages were examined (Additional file 6;Additional file 7). One of these is specific to NB liver(peak N1) and the other is a large region (peak N2) methy-lated in E15 and NB of liver and testis and in AD heart,and demethylated in adult liver and testis. Overall, ourresults indicate extensive tissue and developmental stagespecific methylation in the Hoxa gene cluster that is notassociated with promoter or CpGi regions.DNA methylation within the Protocadherin gene clusterProtocadherins are a large gene family involved in celladhesion and signaling, particularly during neural devel-opment. There are three large clusters totaling morethan 50 genes (alpha, beta and gamma) that span about1 MB on Chromosome 18 in the mouse. Additional file8 shows the MeDIP methylation profile from a portionof the Pcdha and Pchdhg gene clusters. Many of thealternative promoters are associated with CpG islands.Most of the alternative promoters are completelyunmethylated in the adult testis, but some are differen-tially methylated in the liver (Pcdhga3 and Pcdhgb2,peaks 8 and 11; Additional file 8 and Additional file 9).Note that the promoter for Pcdhga6 (peak 13) is methy-lated in the testis but unmethylated in liver, brain, andheart. These results indicate that alternative promoterswithin the Protocadherin gene cluster are targets forDNA methylation.Relationship between T-DMRs/DS-DMRs and geneexpressionTo test whether the T-DMRs and DS-DMRs have anyrelationship with and the expression of their associatedgenes, we identified (from NCBI GEO database) andanalyzed two published expression data sets, which havethe best match to the samples we used in this studywith regard to animal strain, tissue type and develop-mental stage.For T-DMRs, we used GSE9954 [43] designed for theanalysis of 22 different tissues from adult C57BL/6 mice,and we performed a small-scale manual examination, aswell as computational analysis for all T-DMRs. Manualexamination of individual genes revealed a variety of situa-tions, with no obvious consistent pattern, likely due to rea-sons discussed later. Nevertheless, the data from theanalysis of all T-DMRs seems to reveal two general trends:1) Genes containing T-DMRs in non-CpGi promoterstend to have lower expression in the tissue to which theT-DMRs are unique than in other three tissues (Addi-tional file 10 panel A-C); 2) Within the same tissue, genescontaining T-DMRs in non-CpGi promoters have lowerexpression than those containing T-DMRs in either inCpGi promoters or in intragenic CpGi and those contain-ing no T-DMRs (Additional file 10 panel D). The firsttrend is well shown in brain and liver, but less consistentin heart and brain, while the second trend is consistentlyobserved in all four tissues. T-DMRs in CpGi promotersdo not seem to associate with a significantly lower expres-sion except for those from liver, while T-DMRs in intra-genic CpGi do not seem to have a significant impact onTable 3 Overlap among DS-DMRs from different tissuesTissue all DS-DMRs Tissue-Unique Multiple tissues# T-DMR%Brain 1973 1065 908 100Heart 867 533 334 100Liver 3570 2160 1410 100Testis 5469 4099 1370 100Total 11879*/9784** 7857 4022*/1927**#: peak present in 2 or more but not all tissues; *: contains redundancy due tooverlap; The numbers are larger than those in Table 2 due to certain peaksplitting; **: a non-redundant (unique) list of all sites from the four tissues.Liang et al. BMC Genomics 2011, 12:231http://www.biomedcentral.com/1471-2164/12/231Page 7 of 16gene expression either way in any tissue examined. Thetrend for DS-DMRs, based on analysis of a liver develop-ment expression data set (GSE13149 [44]), seems to beless clear except for adult unique DS-DMR, which aremostly T-DMRs as described earlier and show a similarpattern as T-DMRs (data now shown), suggesting thatthere might be some differences between T-DMRs andDS-DMRs in the ways they impact gene expression.DiscussionTotal number of T-DMRs and DS-DMRsMeDIP methylation analysis in conjunction with Nim-bleGen promoter + CpGi microarrays were used toidentify adult tissue specific differentially methylatedregions (T-DMRs) and developmental stage specificDMRs (DS-DMRs). We used relatively conservative cri-teria to identify methylation peaks requiring bothA3A4A1A2A5A6A7A9A10A11A13E15 1E15 2NB1NB2AD1AD2E15 1E15 2NB1NB2AD1AD2E15 1E15 2NB1NB2AD1AD2E15 1E15 2NB1NB2AD1AD2Brain Heart Liver Testis0.0 6.92.0P score color scheme87654321N1chr6:52190631chr6:52087777Peak#N2Hox genelocationFigure 3 DS-DMRs: Heat Map of methyation signal in the Hoxa gene region during development in different tissues. A heat map wasconstructed based on the NimbleGen P scores of probes, in the Hoxa gene cluster region. Probes included in the heat map either have a Pscore ≥2 for both of the duplicated samples or at least one tissue/developmental stage or fall into the promoter region of a Hoxa gene. Eachrow of the color tiles in the heat map represents a NimbleGen probe, while each column represents a sample. The tissue name is labeled at thebottom, while the developmental stage is indicated at the top. The labeled arrows at the right side indicate the position of the TSS anddirection of the Hoxa gene transcription. The T-DMRs and DS-DMRs methylation peaks corresponding to those in Additional file 14b areindicated by the numbers on the left side.Liang et al. BMC Genomics 2011, 12:231http://www.biomedcentral.com/1471-2164/12/231Page 8 of 16independently derived biological replicates to containthe same or overlapping methylation peaks. Other inves-tigators have indicated that MeDIP may miss differen-tially methylated regions in regions of low CpG density[45]. Thus we may underestimate the total number ofT-DMRs and DS-DMRs to some extent. Nevertheless,our study identified almost 5,000 adult T-DMRs and10,000 DS-DMRs that represent a total of over 15,000T-DMRs in four adult tissues (brain, heart, liver, andtestis; Tables 1 and 2). Since tissues are made up ofmany cell types and it is likely that there are cell typespecific epigenetic differences, the methylation peaks weidentified probably correspond to the major cell types inthe tissue, while many methylation peaks correspondingto minor cell types would be missed. Considering thatwe only sampled 4 tissues and 3 developmental stages(E15, NB, AD), and that the analysis does not includemethylation peaks in regions outside promoter andCpGi regions (except in tiling regions), it is quite likelythat there are many additional T-DMRs and DS-DMRsthat were not identified in this study. These resultsdemonstrate that there are numerous DNA methylationdifferences between adult tissues (T-DMRs) and withinthe same tissue at different stages of development (DS-DMRs), indicating that alterations in DNA methylationare a major feature of development.Genomic locations of T-DMRs and DS-DMRsOverall most T-DMRs and DS-DMRs (about 70%) arelocated within non-CpGi promoter regions (Table 1 and2), which is consistent with results of other studies[12,28,30,33,39]. However, we also found that about 30%of T-DMRs (1,440/4,686) and DS-DMRs (3,014/9,336)are located in CpGi regions (i.e. the sum of all CpGi-associated DMRs vs. the total number of DMRs; Table 1and 2). This corresponds to more than 15% of the15,979 CpGi regions on the NimbleGen array and is arelatively large number considering the limited numberof tissues and developmental stages surveyed and otherfactors cited above. In fact, among the 15,979 CpGiregions annotated in UCSC mm8 freeze, a total of 5,523(34.6%) were partially or entirely methylated in one ormore of the 24 samples analyzed in this study (4 tissueseach with duplicated samples for 3 developmentalstages; data not shown). Thus, tissue- and developmen-tal stage-specific CpGi methylation may be a very com-mon event during development.Analysis of methylation in the few tiling regions avail-able on the NimbleGen array made it possible to deter-mine whether there was significant tissue specificmethylation outside of promoter and CpG island regions.Our results indicate that as much as 30% of the T-DMRsare missed by restricting analysis to promoter and CpGiregions (Additional file 3b).The location of methylation peaks differs according totissue distribution. About 50% of the adult commonmethylation sites are within CpGi regions whereas only30% of the T-DMRs are located within CpGi regions(Table 1). Similarly, about 48% of the methylation peaksthat are common among all developmental stages withina tissue are located in CpGi regions (Table 2). The sig-nificance of this difference between common methyla-tion and differential methylation is not presently clear.Possibly this reflects a difference between transientchanges in methylation that are tissue specific and morestable changes in methylation that are common totissues.The distribution of the location of T-DMRs and DS-DMRs in testis is dramatically different from somatictissues. Almost 94% of the adult testis T-DMRs areassociated with non-CpGi promoter regions (Table 1and 2). About 68% of the testis DS-DMRs are associatedwith non-CpGi promoter regions, which represents anaverage of E15, NB, and AD developmental stages.Unlike somatic tissues there appears to be a clear pat-tern shift in the distribution of testis DS-DMRs duringdevelopment. In E15 testis only about 37% of the DS-DMRs are located in non-CpGi promoters whereas thepercentage increases in NB (49%) and AD (86%). Thismay reflect post-natal onset of meiosis and other devel-opmental changes. This general bias of DMRs towardsnon-CpGi promoter vs. CpGi promoter is significantsince the majority (10,915 or 68.5%) of 19,528 promo-ters (all RefSeqs promoters included in the array) areCpGi promoters (based on our criteria region of 1kbflanking each side of transcription start site).T-DMRs correspond to sites of methylation that varybetween tissues, whereas DS-DMRs correspond to sitesof methylation within a single tissue that differ accord-ing to developmental stage. We determined the extentof overlap between DS-DMRs from different tissues bycombining all DS-DMRs for the same tissue and exclud-ing common methylation sites within a tissue (Table 3and Additional file 5a and 5b). We reasoned that thedegree of overlap of DS-DMRs among tissues wouldshed light on the level of tissue specificity of DNAmethylation during development. We found a very lowlevel of overlap among different tissues (1,927/9,784 or13.2%), indicating that most DS-DMRs are tissue-uniqueand tissue-specific (Table 3). Even for DMRs that areshared among all three developmental stages in a tissue(Additional file 5b), almost 50% are tissue unique andalmost all are tissue specific (95.2%). In addition, a largefraction of the tissue specific DS-DMRs were unique toa single tissue (7,857/9,784 or 80%). However, we wouldexpect this percentage to decrease as more tissues areadded to the analysis. These results suggest that almostall DNA methylation in non-repetitive regions is tissueLiang et al. BMC Genomics 2011, 12:231http://www.biomedcentral.com/1471-2164/12/231Page 9 of 16specific. We have previously shown, using RLGS, thatmethylation of some genomic regions containing repeti-tive sequences is also tissue-specific [32].Recent studies [22,25] identified 16,379 T-DMRs in fourhuman adult tissues (brain, liver, spleen, and colon) usinga method termed “Comprehensive High-throughputArrays for Relative Methylation” or CHARM. Althoughthe methods, species and tissues used for this analysiswere different from those presented here, the total numberof T-DMRs is surprisingly similar (15,271 vs. 16,379),emphasizing extensive tissue specific DNA differences inDNA methylation. However, the CHARM analysis ofhuman tissues found that 76% of the T-DMRs werelocated within 2 kb of CpGi regions that were denotedCpGi shores. The array design used in our study limitsanalysis to promoter, CpGi regions, and very limited tiledgenomic regions and would appear to exclude most CpGishores. However, we reason that many CpGi shores maybe within the promoter regions in our studies. For boththe promoter and tiling regions, we observed a slightlyhigher density of DS-DMRs in CpGi shores than in CpGiregions (data not shown). Since at the genome scale, theCpGi shore region is much larger (~6X) than the CpGiregion, we can expect to have a larger number of DS-DMRs in the CpGi shores than in the CpGi, supportingthe conclusion of Irizarry et al [25] in principle that thereare more T-DMRs in CpGi shores.Gene ontology of T-DMRs and DS-DMRsGene ontology (GO) analysis was performed to identifyany common theme among the identified T-DMRs. Inthis analysis, we divided all DMRs associated with genesinto three groups (non-CpGi promoter, CpGi promoter,and intragenic CpGi), since as noted by others [37,38],methylation in different gene locations may impact thegene expression regulation differently. We found moreor less similar enrichment categories for differenttissues, especially among somatic tissues. For bothT-DMRs and DS-DMRs, the most consistent GOenrichment pattern is seen among those in non-CpGipromoters (Additional file 11 and Additional file 12)with enrichment for “membrane proteins, G-proteins,olfactory proteins” among UnMe-AD (DS-DMRs) and“extracellular space/region” for both T-DMRs and DS-DMRS (Me-AD). Apparently, these enrichment patternsreflect a pattern that is observed for all genes with non-CpGi promoters (data not shown). Additional GOenrichment was observed for somatic DS-DMRs inintragenic CpGi regions on “regulation of biological andcellular processes, ion-binding and transport”. This datasupports the notion that intra-genic CpGi methylationtends to promote the expression of the involved gene[46,47], and may participate in gene regulation duringdifferentiation and development.“Demethylation” of previously methylated sites (DS-DMRs) is a common feature of tissue differentiationQuite surprisingly, our studies indicate that many DS-DMRs that are methylated at early stages of development(E15 and NB) are unmethylated in adult tissues (Table 2and Figure 2). This is particularly evident in brain andliver where there are almost 3 times as many DS-DMRsthat become “unmethylated” in adult as become “methy-lated” in adult. This is somewhat contrary to expectationsthat differentiation into adult tissues would reflect pro-moter methylation and silencing of genes not associatedwith the final gene expression pattern. In contrast, it sug-gests that the final gene expression pattern depends onextensive demethylation events during differentiation.Although methylation of gene promoter regions is asso-ciated with gene repression, gene body methylation isassociated with gene expression [37,38,48]. In addition,recent reports indicate that there are extensive allelic dif-ferences in gene expression in human that result fromallelic differences in methylation, due to imprinting orDNA sequence variation [49,50]. Thus, demethylation ofpreviously methylated sites could reflect either increasedor decreased gene expression, depending on site location.Our observations that demethylation is a common fea-ture of tissue differentiation are consistent with therecent report that liver development in human is charac-terized predominantly by demethylation [20].Methylation of ES cellsES cells are pluripotent and are derived from the blasto-meres of the early embryo that are thought to be exten-sively demethylated due to active and passivedemethylation that follows fertilization. Therefore, wedecided to compare the methylation peaks from ES cellswith those found in E15 embryonic tissues (brain, heart,liver, and testis) to determine the extent of methylationdifferences during early stages of tissue differentiation(Additional file 1). We found that the number of E15 T-DMRs (1,769) is less than half of that found in adult(4,686), which is consistent with E15 tissues being gener-ally less methylated than adult. ES cells also had a lowlevel of methylation with 981 T-DMRs. However, some-what surprisingly, ES cells had more methylation peaks(981, Additional file 1) than any of the E15 tissues. Thiswould suggest that many genomic sites that are methy-lated in ES cells become demethylated during early devel-opment. We previously found that many (60%) of theadult T-DMRs identified by RLGS were methylated in EScells, suggesting these were demethylated during tissue dif-ferentiation [32]. It is also possible that some or many ofthe sites that are methylated in ES cells are due to growthin tissue culture. It has previously been shown that tumorcells grown in tissue culture accumulate excessive aberrantmethylation that is unrelated to tumorigenesis [40]. Also,Liang et al. BMC Genomics 2011, 12:231http://www.biomedcentral.com/1471-2164/12/231Page 10 of 16growth of mouse neural progenitor cells in culture afterextended passages results in aberrant methylation [26] andinvestigators found extensive differences in the genomicmethylation patterns of independently isolated human EScell lines [20]. At the present time, it is unclear whethersites of methylation in ES cells is aberrant due to extendedgrowth in culture or whether demethylation during earlydifferentiation in ES cells is an important process as inlater stages of tissue differentiation noted above.Methylation analysis of stem cells revealed extensiveCytosine methylation in a non-CpG context [35,48,51]that was mainly located within gene bodies [35]. Thenon-CpG methylation appears to be mostly confined tostem cells and disappears upon differentiation. Our ana-lysis of mouse ES cells indicated a very low level ofintragenic CpGi methylation. This suggests that most ofthe non-CpG gene body methylation was in non-CpGiintragenic regions that were not on the NimbleGenarray or that our analysis did not resolve non-CpGmethylation.Non-Promoter and non-CpGi methylation in the Hoxagene clusterA 100 Kb region on chromosome 6 that includes theHoxa gene cluster is essentially a tiling array on the Nim-bleGen promoter plus CpGi array. This region includes16 RefSeq genes and 22 UCSC-annotated CpGi regions(Additional file 6). Somewhat surprisingly all 8 methyla-tion peaks were in non-CpGi regions and only one was ina promoter region (Additional file 6 Additional file 7). Inaddition, developmental analysis of Hoxa gene methyla-tion indicated stage-specific methylation differences(Figure 3). These differentially methylated regionsinclude 3’ exons and intron regions. Since these regionsare highly conserved and Hox genes are known to haveimportant roles in development [41,42], these results sug-gest that methylation in non-promoter, non-CpGiregions may have novel, currently undefined roles in reg-ulating development. A recent report also indicates dif-ferential development-associated methylation within Hoxgene clusters in human [48].DNA methylation and alternative promoter useInterrogation of the MeDIP/NimbleGen array data sug-gests that DNA methylation may be associated with alter-native promoter use. Analysis of methylation within theProtocadherin gene clusters on chromosome 18 indicatesextensive methylation within CpGi promoter and non-CpGi promoter regions, particularly within somatic tis-sues (Additional file 8 Additional file 9). Some differentialtissue specific methylation in this region is also noted.Pcdha mRNAs are synthesized by the activation of one ofthe alternative promoters on only one of the two chro-mosomes resulting in monoallelic expression [52].Although the mechanistic basis for this selection isunknown, it is hypothesized that it provides a foundationfor neuron adhesive diversity that is required for complexsynaptic interactions [52]. Recent genome-wide methyla-tion analysis using CHARM [25] also indicates an asso-ciation between tissue specific DNA methylation andalternative transcripts. That study indicated that most tis-sue specific differentially methylated regions were locatedin CpGi shores and that 68% of the shores were within500 bp of alternative promoter sites.The impact of T-DMRs and DS-DMRs on gene expressionTo understand the biological function of T-DMRs andDS-DMRs, one obvious approach is to examine theirimpact on gene expression. Efforts to address this ques-tion at a genome scale is complicated by several factorsthat can obscure correlations.First, one gene may be subjected to DNA methylationin multiple regions with the T-DMR or DS-DMR inquestion being just one of those. Therefore, the level ofgene expression may depend on methylation status inother regions of the gene. Second, existence of multiplesplicing isoforms, particularly those associated with alter-native promoters, as well as the use of multiple expres-sion probes for the same gene makes this one-gene vs.one DMR association analysis very challenging. Last butnot least, DNA methylation is not the only factor affect-ing the gene expression. In other words, the lack of DNAmethylation in one of the promoter region does notnecessarily confer gene activation, as the expression canbe limited by other factors, for instance the lack of therequired transcriptional factor(s). These complicationsmay be responsible for the highly diverse situations weobserved between the occurrence of T-DMRs/DS-DMRsand the expression level of their associated genes on agene-by-gene basis. Despite these complications, our pre-liminary analysis does seem to reveal a few novel insights.First, it appears that T-DMRs are associated with lowergene expression in non-CpGi promoter regions. Second,there may be some differences between T-DMRs andDS-DMRs from earlier developmental stages in the man-ner they impact gene expression. Certainly, extensivedata analysis using more data sets representing more tis-sues, as well as experimental validation is needed to con-firm these observations.ConclusionsOverall, our results indicate that there are extensive dif-ferences in DNA methylation between adult tissues andwithin tissues during development. These results indi-cate that differential DNA methylation is a dynamic pro-cess involving both methylation and demethylationevents. The prominence of demethylation in adult tis-sues of genomic regions that are methylated at earlierLiang et al. BMC Genomics 2011, 12:231http://www.biomedcentral.com/1471-2164/12/231Page 11 of 16stages of development was particularly striking. Eventhough proliferative changes in cell populations mayunderlie many apparent methylation differences duringdevelopment, the results still indicate that extensive dif-ferences in methylation status are associated with devel-opment. Although the majority of the methylationdifferences were associated with non-CpGi promoterregions, there were also many methylation differencesassociated with non-promoter regions that may havenovel roles in regulating development.MethodsGrowth of ES cells and collection of tissues, DNA andRNA preparationsES cells were grown as previously described [32,53]. C57BL/6J mice were purchased from Jackson Laboratory andmaintained less than three generations in the departmentof Laboratory Animal Research facility at Roswell ParkCancer Institute under an approved IACUC protocol.Conventional pair breeding was performed, the plug wasconfirmed and then the male mouse was removed fromthe breeding cage after one day. Pregnant mice were sacri-ficed for embryo collection 15 days later (E15). Embryonicsacs were kept in PBS solution on ice. After the quick con-firmation of development stage, embryos were subjectedto dissection under the microscope (Leica MZ-125). Brain,heart, liver and testis tissues were carefully obtained frommale embryos (verified later, see following). Male newborn (NB) mice were collected within 24 hours after deliv-ery and used for collection of the same tissues. 12 weeksold adult male mice (AD) were also used for adult tissuecollection as described before [32]. Tissues and cells wereimmediately snap-frozen in liquid nitrogen and stored at -80 °C until use.The sex of embryonic mice was determined by using Ychromosome marker and the Zfxy genotyping. PCR wasperformed to confirm sex by using primers (Zfxya-’:AGCTGTTTCATAGTCACAGAACTTAC, Zfy-’-11:CGAATGTGATGACTGTAGGAAGAATC, and Zfx-5’-2:AGAAAGCCATAGAATGCGATGAGTGC). PCR resultsshowing multiple bands indicated male mice, which werethen used in this study. Mixtures of 6 male mouse tissueswere used for DNA isolation for E15 and NB analysissince it was not feasible to obtain enough DNA from a sin-gle E15 or NB animal. Adult tissues were collected fromtwo separate animals. DNA was then prepared from 2separate mixtures (E15 and NB) and 2 separate adults.Genomic DNA from all tissue samples was extractedusing the DNeasy Kit (QIAGEN) according to Qiagenprotocol.Sequenom MassARRAY quantitative methylation analysisSequenom MassARRAY quantitative methylation analysiswas performed as previously described [32] usingprotocols provided by Sequenom (see Additional file 13for a list of primers used). All oligonucleotides were pur-chased from Integrated DNA Technologies (Coralville,IA). Amplification of 1 μl bisulfite-treated DNA (~20 ng/ml) was performed using HotStar Taq Polymerase (Qia-gen) in a 5 μl reaction volume using PCR primers at a200 nM final concentration. PCR amplification was per-formed with the following parameters: 94 °C for 15 minhot start, followed by denaturing at 94 °C for 20 s,annealing at 56 °C for 30 s, extension at 72 °C for 1 minfor 45 cycles, and final incubation at 72 °C for 3 min.PCR products were processed for MassARRAY analysisaccording to the manufacturer’s instructions (SequenomhMC) by the Microarray and Genomics Core Facility atRoswell Park Cancer Institute. In the mass spectrum, therelative amount of methylation can be calculated by com-paring the signal intensity between mass signals ofmethylated and non-methylated template DNA. Eachspot is ionized 50× per second at five different rasterswith all resultant methylation calls performed by the Epi-Typer software v1.0 (Sequenom). All data is transferredand stored in an Oracle database for tabulation. A mini-mum of two independently derived tissue DNAs wereanalyzed.Methylated DNA Immunoprecipitation analysis: MeDIP [33]High-quality genomic DNA isolated from E15, NB andAD mouse tissues, was digested with MseI (TTAA)(New England Biolabs) to produce small fragments (200- 1,000 bp) while keeping CpGi regions intact. Fragmen-ted DNA was heat denatured to produce single-strandedDNA, and a portion of the denatured DNA stored ascontrol (input) DNA. Monoclonal mouse anti 5-methylcytidine ( Eurogentec) was used to immunoprecipitatemethylated DNA fragments. The immune complexeswere captured with Protein A Agarose beads (Invitro-gen). Complexes were washed to remove nonspecificallybound material. Following elution of bound complexes,ethanol precipitation, and resuspension of MeDIP DNA,a small aliquot of DNA and control input DNA wereused to amplify a known methylated DNA region and aknown unmethylated DNA region by real-time PCR.The primers of Pst4, Pst6 and Pst21 that were identifiedby RLGS previously [32] were used in enrichment mea-surements according to the manufacturer’s instructions(NimbelGen). Ratios between the MeDIP and inputvalues are defined and normalized against a knownunmethylated control sequence. After validating theenrichment of MeDIP DNA, MeDIP DNA and controlinput DNA were amplified by whole-genome amplifica-tion kit (Sigma Aldrich), followed by purification (QIA-GEN Quick PCR Purification Kit) and then sent toNimbleGen for Microarray hybridization according totheir standard protocol for the array.Liang et al. BMC Genomics 2011, 12:231http://www.biomedcentral.com/1471-2164/12/231Page 12 of 16Data analysisRaw data processing and methylation peak identificationThe first stage of raw data processing was performed byNimbleGen as part of its commercial service (Detaileddescription is available at http://www.nimblegen.com).In brief, it involves the following 4 steps. 1) Signalextraction: Signal intensity data is extracted from thescanned images of each array using NimbleScan. 2) Cal-culation of log2 ratio: For each feature on the array, acorresponding scaled log2-ratio representing the ratio ofthe input signals for the experimental and test samplesthat were co-hybridized to the array is computed andscaled to center the ratio data around zero. Scaling isperformed by subtracting the bi-weight mean for thelog2-ratio values for all features on the array from eachlog2-ratio value. 3) Calculation of P score: From thescaled log2-ratio data, a fixed-length window (750 bp) isplaced around each consecutive probe and the one-sided Kolmogorov-Smirnov (KS) test is applied to deter-mine whether the probes are drawn from a significantlymore positive distribution of intensity log-ratios thanthose in the rest of the array. The resulting score foreach probe, called P score, is the -log10 p-value from thewindowed KS test around that probe. 4) Identificationof methylation peaks: This is performed by searching forregions containing at least 2 consecutive probes with aP score ≥ 2. Peaks within 500 bp of each other aremerged. The processed data (P Scores) and raw data(log2 ratios) for all 24 involved samples have beendeposited in GEO at NCBI under accession# GSE21415.Identification of methylation peaks shared by or unique toa given set of samplesThis part was mostly done using an in-house bioinfor-matics tool based on the following criteria and definition:1) A methylation peak of a sample is the methylated regionshared between the two biological replicates of the samesample. This requirement applies to all categories ofmethylation peaks; 2) A methylation peak unique to a sam-ple or a group of samples is a methylated region that onlyappeared in the sample or group. It was further examinedto ensure that this region does not overlap with any rawpeaks (those of individual replicates from NimbleScan) ofother individual samples; 3) A peak shared by one or morebut not all samples in a group; 4) A methylation peak com-mon to a group of samples. In identifying these methyla-tion peaks, it is often necessary to break the largeroverlapping raw peaks into several sub-peaks such that onepart of an original peak may be identified as a commonpeak and another part as a unique peak.Defining the location category of DMRsTo define the location of a methylation peak in contextof a gene, we used the annotation data of promotersand CpGi regions of mouse genome provided byNimbleGen, which is in turn based on UCSC February2006 (mm8) genome freeze for annotation of CpGiregions and mRNAs and their genome coordinates.Using an in-house PERL script, each peak is assignedinto one of the following categories: 1) CpGi promoter,located within 500 bp distance to a CpGi and is withinan annotated promoter (1 kb flanking each side of thetranscription start site or TSS); 2) non-CpGi promoter,located within a promoter without CpGi or more than500 bp away from any CpGi in the promoter; 3) Intra-genic CpGi, located within 500 bp of a CpGi that locatesin a gene (excluding the promoter as defined above); 4)Intergenic CpGi, located within 500 bp from a CpGithat is more than 2 kb from any annotated genes. Theassignment is done in the above order of the four cate-gories such that a given DMR belongs to only one cate-gory with the CpGi promoter having a higher priorityover the rest of the categories and so on. For DMRs intiling regions, two additional categories were added asnon-CpGi intergenic and non-CpGi intragenic. DMRs intiling regions were identified by collecting all DMRslocated within the tiling regions, which include Chr6:51,984,676-52,408,211, Chr7: 142,162,309-143,552,631and Chr17: 12,525,776-12,614,350.Gene ontology analysis and other statistical analysesFor DMRs associated with a gene (including CpGi pro-moter, non-CpGi promoter, and Intragenic CpGi), weexamined whether a statistically significant enrichment ofgene ontology - a given gene list can be observed usingthe DAVID tool [54]. A significant enrichment requires aBenjamini adjusted P value no more than 0.05. Scatterplots were generated using R http://www.r-project.org/and Heat maps of P scores were generated using MeVhttp://www.tm4.org/mev/[55].Assessment of the MeDIP data quality and validation ofMeDIP methylation peaksTo evaluate the overall quality of the MeDIP data, weplotted the log2 ratios (methylation signals between the IPDNA and the input DNA) and P score of all probesbetween biological replicates representing different sam-ples of the same tissues from different animals. The scatterplots of log 2 ratios as shown in Additional file 14 indi-cates that a good level of consistency was seen among bio-logical replicates for all examined tissues (Additional file4), considering these represented both biological and tech-nical replicates (Pearson Coefficient R = 0.68 to 0.88 withover half being > 0.8). The Pearson Coefficient betweenreplicates is improved in all cases, ranging from 0.75 to0.94, when P score was used (Additional file 4), on whichthe peak call was based, indicating some sporadic noisewas removed by considering the signal level of neighbor-ing probes in a larger window.Liang et al. BMC Genomics 2011, 12:231http://www.biomedcentral.com/1471-2164/12/231Page 13 of 16We previously identified T-DMRs using RestrictionLandmark Genomic Scanning (RLGS) [19,32], and someof these RLGS T-DMRs were used as internal controlsfor the quality of IP DNA prior to the array hybridiza-tion. Quantitative PCR analysis of input and IP DNAsamples indicated that the IP DNA from the appropriatetissue was enriched for DNA sequences previouslyshown to have tissue specific DNA methylation (pre-sence or absence of methylated DNA fragments at Pst6and Pst21 were confirmed prior to the hybridization;data not shown). Additional file 15 shows the results ofMeDIP methylation analysis in the regions of four pre-viously identified RLGS T-DMRs (Pst4, Pst10, Pvu4 andPvu5) that were confirmed by Sequenom MassArrayquantitative methylation analysis [32]. These results areconsistent with both RLGS analysis and the Sequenomanalysis of bisulfite treated DNA [32]. Among the 68RLGS-TDMRs, 38 overlap with the probes on the array(allowing a distance of 750 bp the window size we usedfor peak identification), of which 27 overlap with theMeDIP DS-DMRs or T-DMRs. Among the 11 notpicked up by MeDIP, 5 are in the repetitive regions andwould not be included on the array. Therefore, the sen-sitivity level of our MeDIP in reference to RLGS is 27/33 or 82%.Sequenom MassArray quantitative methylation analysisof bisulfite treated DNA was performed as an additionalmeans to confirm DNA methylation peaks obtained byMeDIP/NimbleGen Array for both selected and randomlychosen MeDIP methylated regions. The Sequenom Mas-sArray results were consistent with the MeDIP results in19 of 23 peaks that were evaluated (18 randomly chosenand 5 selected peaks). An example of the Sequenom Mas-sArray results for a CpGi promoter region is shown inFigure 1, and for a randomly selected non-CpGi promoterregion (Additional file 16) indicating the two methods arein agreement.We also interrogated MeDIP data within imprintedregions that are known to be methylated in somatic tis-sues. Almost the entire imprinted region on Chr 7 thatincludes the Kcnq1 gene is represented in a tiling array(1,390 kb) on the NimbleGen promoter + CpGi array. A40 kb portion of the Kcnq1 that contains KvDMR isshown in Additional file 2. The KvDMR is an imprintedcontrol region of approximately 2 Kb (boxed region)that is maternally methylated and has previously beenshown to be unmethylated in sperm [56]. Consistentwith these earlier results, the MeDIP/NimbleGen arrayresults clearly indicate that DNA from the testis is notmethylated in this region. Similarly, the imprinting con-trol region associated with Igf2r is also unmethylated intestis [57] (Additional file 17). Since only the maternalallele is methylated in these imprinted regions, MeDIPis capable of detecting methylation in a single allele.Analysis of relationship between T-DMRs/DS-DMRs andgene expressionTwo gene expression data sets, GSE9954 and GSE13149,that are deemed suitable for our purpose were identifiedand retrieved from the NCBI GEO database. ForGSE9954, we used the normalized expression values pro-vided in GEO by the authors and performed the distribu-tion pattern and statistical analysis of expression for genesassociated with all T-DMRs from each tissue. Specifically,for T-DMRs from each of the 3 location categories, i.e.,the CpGi promoter, the non-CpGi promoter and intra-genic CpGi, in a tissue, we compared the expression leveland distribution pattern of their associated genes in thattissue with that of the same set of genes in three other tis-sues, as well as gene associated T-DMRs in other two loca-tion categories, based on box plot and t-test using the Rpackage. For GSE13149, the raw Cel files were down-loaded from GEO for samples at E15.5, Day0 and adultmice, which presumably correspond to our E15, new born(NB) and adult samples, and were processed for gcRMAand quantile normalization using the BioConductor Rpage. The expression data in association with liver DS-DMRs was analyzed in a manner similar to T-DMRs.Additional materialAdditional file 1: E15 and ES T-DMRs.Additional file 2: MeDIP/NimbleGen Promoter + CpGi Array (Tilingregion): Methylation analysis of KvDMR region. The scaled log2 ratio ofthe 40 Kb region on chromosome 7 near KvDMR is shown. The numbers onthe top indicate the genomic position. The rectangle indicates the positionof the KvDMR that includes two CpGi regions. Two independent tissueswere taken from different mice. Previous studies indicate that the KvDMR ismethylated in somatic tissue and unmethylated in sperm [57].Additional file 3: T-DMRs in tiling regions.Additional file 4: Pearson Coefficient and common peak numberbetween replicates.Additional file 5: a: Overlaps between different tissues for DS-DMRswith testis excluded; b:Overlaps between tissues for DMRs sharedby all developmental stages in a tissue.Additional file 6: T-DMRs within the Hoxa gene cluster. The MeDIPmethylation profile for the Hoxa gene cluster within the 52.08 to 52.20 Mbregion on chromosome 6 is shown. The locations of CpGi regions, Hoxagene transcripts from a1 to a13 and transcription direction are indicated.The numbers on the bottom from 1 to 8 indicate the position ofmethylation peaks (represented by the log2 ratio) corresponding to thoselisted in Additional file 7.Additional file 7: Methylation peaks locations in the Hoxa genecluster.Additional file 8: DNA methylation and alternative promoters forPcdha genes. Methylation profile for Pcdha genes is shown in a similarmanner as in Additional file 2. See also Additional file 9.Additional file 9: Methylation peaks locations in the Protocadheringene cluster.Additional file 10: Relationship between T-DMRs and expression oftheir associated genes. The expression of the genes associated withliver unique T-DMRs were compared to that of the same gene set in thethree other tissues broken down into the CpGi promoter (CpGi_P), non-CpGi promoter (NCpGi_P) and intra-genic CpGi (Intra_CpGi) groups usingLiang et al. BMC Genomics 2011, 12:231http://www.biomedcentral.com/1471-2164/12/231Page 14 of 16boxplots (panel A, B and C, respectively). Comparison was also made forT-DMRs associated genes across different location groups within liver(Panel D). The p values of pairwise t-tests between liver and each of theother tissue (panels A-C) and between each pair of location groups(panel C) are provided on the right side of the boxplots with those nogreat than 0.05 shown in bold font. “NDMR” in panel C refers to genesassociated with all genes not associated with liver unique T-DMRs.Additional file 11: Gene ontology for T-DMRs in adult.Additional file 12: Gene ontology for DS-DMRs subjected tomethylation or demethylation in adult.Additional file 13: A list of Sequenom MassARRAY primers used forValidation.Additional file 14: Scatter plots of log2 ratios of two replicates forbrain tissues at E15, NB, and AD stages. Log2 ratio of all probes forthe two replicate samples of brain at all stages were plotted. In eachplot, data points in red rectangular area are those showing log2 ratio ≥ 1(i.e. ratio ≥2) in both samples. Pearson Coefficient and the commonmethylation peak number are provided on the top of each plot.Additional file 15: MeDIP/NimbleGen Promoter + CpGi Array:Methylation analysis of RLGS T-DMR loci. The MeDIP methylationprofile including scaled log2 ratio and methylation peaks of four RLGS T-DMR loci are shown (Pst4, Pst10, Pvu4 and Pvu5). The loci were identifiedas T-DMRs by RLGS [32]. The CpG island, genomic location, relevant geneand the direction of each transcript are indicated. The log2 ratio is theratio of signals for the input and immunoprecipitated DNA test samplesthat were co-hybridized to the array. RLGS T-DMRs were also confirmedby Sequenom MassARRAY ([32], and data not shown).Additional file 16: Comparison of methylation data between MeDIP/NimbleGen Promoter + CpGi Array and Sequenom MassARRAY: NonCpGi promoter region (randomly selected). Data similar to thatpresented in Figure 1 for a randomly selected locus, the non-CpGipromoter region of Gm1070 that is methylated in liver and ES, but not intestis.Additional file 17: MeDIP/NimbleGen Promoter + CpGi Array (Tilingregion): Methylation analysis of Igf2r imprint control region. TheMeDIP methylation profile of a region on chromosome 17 that includesIgf2r and BC009123 is shown. The numbers on the top indicate thegenomic position. The CpGi in the rectangle is located close to thepromoter region of BC009123 and the second intron of Igf2r. Methylationwithin the CpGi is associated with repression of transcription of Airnantisense RNA located somewhat downstream of TSS for BC009123. Theimprinting control region (rectangle) is monoallelically methylated insomatic tissue and unmethylated in testis [57].AcknowledgementsThis research was in part supported by the National Cancer Institute GrantCA102423 (WAH) and the National Cancer Institute Core Center GrantCA16056 (Roswell Park Cancer Institute), National Institute of EnvironmentalHealth Science Grant ES012249 (HN), Academic frontier project for 2006from MEXT (HN) and grants from the Canada Research Chair program,Canadian Foundation of Innovation, Ontario Ministry of Research and BrockUniversity (PL). We thank Arlene Morrow for the preparation of some of theDNAs used in these studies, and Michael Higgins and Dominic Smiraglia forsuggestions and critical reading of the manuscript.Author details1Department of Biological Sciences, Brock University, St. Catharines, Ontario,Canada. 2Department of Molecular and Cellular Biology, Roswell Park CancerInstitute, Buffalo, NY, USA. 3Department of Pharmacology and Therapeutics,Roswell Park Cancer Institute, Buffalo, NY, USA. 4Department of CancerGenetics, Roswell Park Cancer Institute, Buffalo, NY, USA. 5Department ofZoology, North-Eastern Hill University, Umshing Mawkynroh, Shillong,Meghalaya, India. 6Department of Bioinformatics, University of BritishColumbia, Vancouver, British Columbia, Canada. 7Department ofBiochemistry, State University of New York at Buffalo, Buffalo, NY, USA.8Division of Cancer Genetics, Department of Advanced Medical Science,Nihon University School of Medicine, Tokyo, Japan. 9Division of CancerGenetics, Chiba Cancer Center, Research Institute, Chiba, Japan.Authors’ contributionsPL: performed most of the bioinformatics data analysis and interpretationand contributed in writing the manuscript; FS: performed most of themolecular analysis including preparation for MeDIP and MassARRAY analysisand contributed in writing, SG: performed validation experiments, EM:performed data analysis; MQ: developed the computational tools foridentifying methylation peaks shared by or unique to different samplegroups, SM: performed ES cell and DNA preparation and validationexperiments, KF and IJ: performed tissue preparation, HN: performedsupervision of most of molecular biological analysis and contributed inwriting the manuscript, WH: performed most of writing and supervision ofboth bioinformatics and molecular analysis. All authors read and approvedthe final manuscript.Received: 25 June 2010 Accepted: 11 May 2011 Published: 11 May 2011References1. Bird AP: DNA methylation and the frequency of CpG in animal DNA.Nucleic Acids Res 1980, 8:1499-1504.2. Antequera F, Bird A: Number of CpG islands and genes in human andmouse. Proc Natl Acad Sci USA 1993, 90:11995-11999.3. Yoder JA, Walsh CP, Bestor TH: Cytosine methylation and the ecology ofintragenomic parasites. Trends Genet 1997, 13:335-340.4. Deng J, Shoemaker R, Xie B, Gore A, LeProust EM, Antosiewicz-Bourget J,Egli D, Maherali N, Park IH, Yu J, Daley GQ, Eggan K, Hochedlinger K,Thomson J, Wang W, Gao Y, Zhang K: Targeted bisulfite sequencingreveals changes in DNA methylation associated with nuclearreprogramming. Nat Biotechnol 2009, 27:353-360.5. Kafri T, Ariel M, Brandeis M, Shemer R, Urven L, McCarrey J, Cedar H,Razin A: Developmental pattern of gene-specific DNA methylation in themouse embryo and germ line. Genes Dev 1992, 6:705-714.6. Monk M, Boubelik M, Lehnert S: Temporal and regional changes in DNAmethylation in the embryonic, extraembryonic and germ cell lineagesduring mouse embryo development. Development 1987, 99:371-382.7. Reik W, Dean W, Walter J: Epigenetic reprogramming in mammaliandevelopment. Science 2001, 293:1089-1093.8. Ehrlich M: Expression of various genes is controlled by DNA methylationduring mammalian development. J Cell Biochem 2003, 88:899-910.9. Futscher BW, Oshiro MM, Wozniak RJ, Holtan N, Hanigan CL, Duan H,Domann FE: Role for DNA methylation in the control of cell type specificmaspin expression. Nat Genet 2002, 31:175-179.10. Riggs AD: X inactivation, differentiation, and DNA methylation. CytogenetCell Genet 1975, 14:9-25.11. Walsh CP, Chaillet JR, Bestor TH: Transcription of IAP endogenous retrovirusesis constrained by cytosine methylation. Nat Genet 1998, 20:116-117.12. Weber M, Hellmann I, Stadler MB, Ramos L, Paabo S, Rebhan M,Schubeler D: Distribution, silencing potential and evolutionary impact ofpromoter DNA methylation in the human genome. Nat Genet 2007,39:457-466.13. Holliday R, Pugh JE: DNA modification mechanisms and gene activityduring development. Science 1975, 187:226-232.14. Li E, Bestor TH, Jaenisch R: Targeted mutation of the DNA methyltransferasegene results in embryonic lethality. Cell 1992, 69:915-926.15. Okano M, Bell DW, Haber DA, Li E: DNA methyltransferases Dnmt3a andDnmt3b are essential for de novo methylation and mammaliandevelopment. Cell 1999, 99:247-257.16. Hatada I, Hayashizaki Y, Hirotsune S, Komatsubara H, Mukai T: A genomicscanning method for higher organisms using restriction sites aslandmarks. Proc Natl Acad Sci USA 1991, 88:9523-9527.17. Kawai J, Hirotsune S, Hirose K, Fushiki S, Watanabe S, Hayashizaki Y:Methylation profiles of genomic DNA of mouse developmental braindetected by restriction landmark genomic scanning (RLGS) method.Nucleic Acids Res 1993, 21:5604-5608.18. Shiota K, Kogo Y, Ohgane J, Imamura T, Urano A, Nishino K, Tanaka S,Hattori N: Epigenetic marks by DNA methylation specific to stem, germand somatic cells in mice. Genes Cells 2002, 7:961-969.19. Song F, Smith JF, Kimura MT, Morrow AD, Matsuyama T, Nagase H,Held WA: Association of tissue-specific differentially methylated regionsLiang et al. BMC Genomics 2011, 12:231http://www.biomedcentral.com/1471-2164/12/231Page 15 of 16(TDMs) with differential gene expression. Proc Natl Acad Sci USA 2005,102:3336-3341.20. Brunner AL, Johnson DS, Kim SW, Valouev A, Reddy TE, Neff NF, Anton E,Medina C, Nguyen L, Chiao E, Oyolu CB, Schroth GP, Absher DM, Baker JC,Myers RM: Distinct DNA methylation patterns characterize differentiatedhuman embryonic stem cells and developing human fetal liver. GenomeRes 2009, 19:1044-1056.21. Ching TT, Maunakea AK, Jun P, Hong C, Zardo G, Pinkel D, Albertson DG,Fridlyand J, Mao JH, Shchors K, Weiss WA, Costello JF: Epigenome analysesusing BAC microarrays identify evolutionary conservation of tissue-specific methylation of SHANK3. Nat Genet 2005, 37:645-651.22. Doi A, Park IH, Wen B, Murakami P, Aryee MJ, Irizarry R, Herb B, Ladd-Acosta C, Rho J, Loewer S, Miller J, Schlaeger T, Daley GQ, Feinberg AP:Differential methylation of tissue- and cancer-specific CpG island shoresdistinguishes human induced pluripotent stem cells, embryonic stemcells and fibroblasts. Nat Genet 2009, 41:1350-1353.23. Illingworth R, Kerr A, Desousa D, Jorgensen H, Ellis P, Stalker J, Jackson D,Clee C, Plumb R, Rogers J, Humphray S, Cox T, Langford C, Bird A: A novelCpG island set identifies tissue-specific methylation at developmentalgene loci. PLoS Biol 2008, 6:e22.24. Khulan B, Thompson RF, Ye K, Fazzari MJ, Suzuki M, Stasiek E, Figueroa ME,Glass JL, Chen Q, Montagna C, Hatchwell E, Selzer RR, Richmond TA,Green RD, Melnick A, Greally JM: Comparative isoschizomer profiling ofcytosine methylation: the HELP assay. Genome Res 2006, 16:1046-1055.25. Irizarry RA, Ladd-Acosta C, Wen B, Wu Z, Montano C, Onyango P, Cui H,Gabo K, Rongione M, Webster M, Ji H, Potash JB, Sabunciyan S,Feinberg AP: The human colon cancer methylome shows similar hypo-and hypermethylation at conserved tissue-specific CpG island shores.Nat Genet 2009, 41:178-186.26. Meissner A, Mikkelsen TS, Gu H, Wernig M, Hanna J, Sivachenko A, Zhang X,Bernstein BE, Nusbaum C, Jaffe DB, Gnirke A, Jaenisch R, Lander ES:Genome-scale DNA methylation maps of pluripotent and differentiatedcells. Nature 2008, 454:766-770.27. Oakes CC, La SS, Smiraglia DJ, Robaire B, Trasler JM: A uniqueconfiguration of genome-wide DNA methylation patterns in the testis.Proc Natl Acad Sci USA 2007, 104:228-233.28. Rakyan VK, Down TA, Thorne NP, Flicek P, Kulesha E, Graf S, Tomazou EM,Backdahl L, Johnson N, Herberth M, Howe KL, Jackson DK, Miretti MM,Fiegler H, Marioni JC, Birney E, Hubbard TJ, Carter NP, Tavare S, Beck S: Anintegrated resource for genome-wide identification and analysis ofhuman tissue-specific differentially methylated regions (tDMRs). GenomeRes 2008, 18:1518-1529.29. Rauch T, Li H, Wu X, Pfeifer GP: MIRA-assisted microarray analysis, a newtechnology for the determination of DNA methylation patterns,identifies frequent methylation of homeodomain-containing genes inlung cancer cells. Cancer Res 2006, 66:7939-7947.30. Schilling E, Rehli M: Global, comparative analysis of tissue-specificpromoter CpG methylation. Genomics 2007, 90:314-323.31. Shen L, Kondo Y, Guo Y, Zhang J, Zhang L, Ahmed S, Shu J, Chen X,Waterland RA, Issa JP: Genome-wide profiling of DNA methylation revealsa class of normally methylated CpG island promoters. PLoS Genet 2007,3:2023-2036.32. Song F, Mahmood S, Ghosh S, Liang P, Smiraglia DJ, Nagase H, Held WA:Tissue specific differentially methylated regions (TDMR): Changes inDNA methylation during development. Genomics 2009, 93:130-139.33. Weber M, Davies JJ, Wittig D, Oakeley EJ, Haase M, Lam WL, Schubeler D:Chromosome-wide and promoter-specific analyses identify sites ofdifferential DNA methylation in normal and transformed human cells.Nat Genet 2005, 37:853-862.34. Yagi S, Hirabayashi K, Sato S, Li W, Takahashi Y, Hirakawa T, Wu G, Hattori N,Hattori N, Ohgane J, Tanaka S, Liu XS, Shiota K: DNA methylation profile oftissue-dependent and differentially methylated regions (T-DMRs) inmouse promoter regions demonstrating tissue-specific gene expression.Genome Res 2008, 18:1969-1978.35. Lister R, Pelizzola M, Dowen RH, Hawkins RD, Hon G, Tonti-Filippini J, Nery JR,Lee L, Ye Z, Ngo QM, Edsall L, Antosiewicz-Bourget J, Stewart R, Ruotti V,Millar AH, Thomson JA, Ren B, Ecker JR: Human DNA methylomes at baseresolution show widespread epigenomic differences. Nature 2009,462:315-322.36. Kitamura E, Igarashi J, Morohashi A, Hida N, Oinuma T, Nemoto N, Song F,Ghosh S, Held WA, Yoshida-Noro C, Nagase H: Analysis of tissue-specificdifferentially methylated regions (TDMs) in humans. Genomics 2007,89:326-337.37. Ball MP, Li JB, Gao Y, Lee JH, LeProust EM, Park IH, Xie B, Daley GQ,Church GM: Targeted and genome-scale strategies reveal gene-bodymethylation signatures in human cells. Nat Biotechnol 2009, 27:361-368.38. Hellman A, Chess A: Gene body-specific methylation on the active ×chromosome. Science 2007, 315:1141-1143.39. Eckhardt F, Lewin J, Cortese R, Rakyan VK, Attwood J, Burger M, Burton J, Cox TV,Davies R, Down TA, Haefliger C, Horton R, Howe K, Jackson DK, Kunde J,Koenig C, Liddle J, Niblett D, Otto T, Pettett R, Seemann S, Thompson C, West T,Rogers J, Olek A, Berlin K, Beck S: DNA methylation profiling of humanchromosomes 6, 20 and 22. Nat Genet 2006, 38:1378-1385.40. Smiraglia DJ, Rush LJ, Fruhwald MC, Dai Z, Held WA, Costello JF, Lang JC,Eng C, Li B, Wright FA, Caligiuri MA, Plass C: Excessive CpG islandhypermethylation in cancer cell lines versus primary humanmalignancies. Hum Mol Genet 2001, 10:1413-1419.41. Garcia-Fernandez J: Hox, ParaHox, ProtoHox: facts and guesses. Heredity2005, 94:145-152.42. Pearson JC, Lemons D, McGinnis W: Modulating Hox gene functionsduring animal body patterning. Nat Rev Genet 2005, 6:893-904.43. Thorrez L, Van DK, Tranchevent LC, Van LL, Engelen K, Marchal K, Moreau Y,Van M, Schuit F: Using ribosomal protein genes as reference: a tale ofcaution. PLoS One 2008, 3:e1854.44. Li T, Huang J, Jiang Y, Zeng Y, He F, Zhang MQ, Han Z, Zhang X: Multi-stage analysis of gene expression and transcription regulation in C57/B6mouse liver development. Genomics 2009, 93:235-242.45. Irizarry RA, Ladd-Acosta C, Carvalho B, Wu H, Brandenburg SA, Jeddeloh JA,Wen B, Feinberg AP: Comprehensive high-throughput arrays for relativemethylation (CHARM). Genome Res 2008, 18:780-790.46. Murrell A, Heeson S, Bowden L, Constancia M, Dean W, Kelsey G, Reik W:An intragenic methylated region in the imprinted Igf2 gene augmentstranscription. EMBO Rep 2001, 2:1101-1106.47. Rauch TA, Wu X, Zhong X, Riggs AD, Pfeifer GP: A human B cellmethylome at 100-base pair resolution. Proc Natl Acad Sci USA 2009,106:671-678.48. Laurent L, Wong E, Li G, Huynh T, Tsirigos A, Ong CT, Low HM, KinSung KW, Rigoutsos I, Loring J, Wei CL: Dynamic changes in the humanmethylome during differentiation. Genome Res 2010, 20:320-331.49. Schalkwyk LC, Meaburn EL, Smith R, Dempster EL, Jeffries AR, Davies MN,Plomin R, Mill J: Allelic skewing of DNA methylation is widespread acrossthe genome. Am J Hum Genet 2010, 86:196-212.50. Kerkel K, Spadola A, Yuan E, Kosek J, Jiang L, Hod E, Li K, Murty VV,Schupf N, Vilain E, Morris M, Haghighi F, Tycko B: Genomic surveys bymethylation-sensitive SNP analysis identify sequence-dependent allele-specific DNA methylation. Nat Genet 2008, 40:904-908.51. Ramsahoye BH, Biniszkiewicz D, Lyko F, Clark V, Bird AP, Jaenisch R: Non-CpG methylation is prevalent in embryonic stem cells and may bemediated by DNA methyltransferase 3a. Proc Natl Acad Sci USA 2000,97:5237-5242.52. Hirayama T, Yagi T: The role and expression of the protocadherin-alphaclusters in the CNS. Curr Opin Neurobiol 2006, 16:336-342.53. Forrester LM, Bernstein A, Rossant J, Nagy A: Long-term reconstitution ofthe mouse hematopoietic system by embryonic stem cell-derived fetalliver. Proc Natl Acad Sci USA 1991, 88:7514-7517.54. Dennis G, Sherman BT, Hosack DA, Yang J, Gao W, Lane HC, Lempicki RA:DAVID: Database for Annotation, Visualization, and Integrated Discovery.Genome Biol 2003, 4:3.55. Saeed AI, Bhagabati NK, Braisted JC, Liang W, Sharov V, Howe EA, Li J,Thiagarajan M, White JA, Quackenbush J: TM4 microarray software suite.Methods Enzymol 2006, 411:134-193.56. Engemann S, Strodicke M, Paulsen M, Franck O, Reinhardt R, Lane N,Reik W, Walter J: Sequence and functional comparison in the Beckwith-Wiedemann region: implications for a novel imprinting centre andextended imprinting. Hum Mol Genet 2000, 9:2691-2706.57. Stoger R, Kubicka P, Liu C-G, Kafri T, Razin A, Cedar H, Barlow P: Maternal-specific methylation of the imprinted mouse lgf2r locus identifies theexpressed locus as carrying the imprinting signal. Cell 1993, 73:61-71.doi:10.1186/1471-2164-12-231Cite this article as: Liang et al.: Genome-wide survey reveals dynamicwidespread tissue-specific changes in DNA methylation duringdevelopment. BMC Genomics 2011 12:231.Liang et al. BMC Genomics 2011, 12:231http://www.biomedcentral.com/1471-2164/12/231Page 16 of 16

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:
https://iiif.library.ubc.ca/presentation/dsp.52383.1-0223776/manifest

Comment

Related Items