UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A genome-wide linkage scan and targeted family-based association analysis of dyslexia Ryan, Jane 2008

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

Item Metadata


24-ubc_2008_fall_ryan_jane.pdf [ 1.11MB ]
JSON: 24-1.0066433.json
JSON-LD: 24-1.0066433-ld.json
RDF/XML (Pretty): 24-1.0066433-rdf.xml
RDF/JSON: 24-1.0066433-rdf.json
Turtle: 24-1.0066433-turtle.txt
N-Triples: 24-1.0066433-rdf-ntriples.txt
Original Record: 24-1.0066433-source.json
Full Text

Full Text

   A GENOME-WIDE LINKAGE SCAN  AND TARGETED FAMILY-BASED ASSOCIATION ANALYSIS  OF DYSLEXIA   by   JANE RYAN  B.Sc., The University of British Columbia, 2003   A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE  in  THE FACULTY OF GRADUATE STUDIES  (Medical Genetics)             THE UNIVERSITY OF BRITISH COLUMBIA  (Vancouver)  June 2008   © Jane Ryan, 2008  ii  ABSTRACT As a specific reading disability with a neurobiological origin, developmental dyslexia is distinct from reading difficulties due to sensory impairments in vision or hearing. The disability is commonly attributed to a core deficit in phonological processing, the understanding of how phonemes, syllables and words are used in a language. Dyslexia is a complex genetic disorder with a strong genetic component; nine susceptibility loci (DYX1-9) have been identified with eight other dyslexia linkages lacking gene symbols also reported. The statistical methods of linkage and association were employed to investigate the genetic susceptibility for phonological coding dyslexia (PCD), a common form of dyslexia characterized by difficulties in single word decoding and resulting from deficits in phonological processing. A genome-wide non-parametric linkage (NPL) study and four targeted fine-mapping family-based association studies were performed to locate the genes predisposing to PCD in 101 Canadian families with multiple affected members. The NPL scan identified suggestive evidence for linkage with PCD at the two novel regions 16p12 and 4q12-q13, and provided independent confirmation of linkage to the well-replicated DYX3 locus (at 2p21). Some support for linkage was noted at a further five regions previously linked to dyslexia, while no linkage was detected at five other reportedly-linked regions, in particular, no linkage to DYX2 (6p22.2). Four regions (16p12, 2p21, 4q12-q13 and 6p22.2) were tested for association with PCD in 83 trios, a subset of the 101 families, using the transmission disequilibrium test (TDT) and the affected family-based controls (AFBAC) test. Association was detected in each of the three PCD-linked regions in the NPL scan; none of the tested marker alleles was associated with PCD in the 6p22.2 region. Four candidate genes were identified, two of which belong to the same gene family, with a possible role in the neurodevelopmental mechanisms underlying reading.  iii  TABLE OF CONTENTS ABSTRACT ....................................................................................................................................... ii TABLE OF CONTENTS ...................................................................................................................... iii LIST OF TABLES ............................................................................................................................... v LIST OF FIGURES ............................................................................................................................. vi LIST OF ABBREVIATIONS ................................................................................................................ vii ACKNOWLEDGEMENTS .................................................................................................................... ix CHAPTER I INTRODUCTION / REVIEW OF THE LITERATURE .................................................... 1 1.1 Developmental Dyslexia: Phenotype and Prevalence ................................................. 1 1.2 Theories of Developmental Dyslexia .......................................................................... 3 1.3 Neurological Studies of Dyslexia ............................................................................... 5 1.4 Genetics of Dyslexia ................................................................................................... 6 1.4.1 Heritability and Twin Studies ............................................................................. 6 1.4.2 Linkage and Linkage Disequilibrium Studies ..................................................... 7 1.5 Complex Trait Gene Mapping .................................................................................... 9 1.6 Aim ........................................................................................................................... 13 CHAPTER 2 METHODS .......................................................................................................... 17 2.1 Study Design ............................................................................................................. 17 2.2 Phenotype .................................................................................................................. 17 2.3 Subjects ..................................................................................................................... 18 2.3.1 Selection of Subjects for Array-Genotyping..................................................... 19 2.3.2 Selection of Subjects for Linkage Analysis ...................................................... 19 2.3.3 Selection of Subjects for Association Analyses ................................................ 20 2.4 Storage of Pedigree and Phenotype data ................................................................... 21 2.5 Array-Genotyping for SNP Markers ......................................................................... 21 2.6 Quality Control Measures ......................................................................................... 22 2.6.1 Initial Removal of Markers that Failed QC Thresholds .................................... 23 2.6.2 Resolution of a Sample Mix-up ........................................................................ 23 2.6.3 Removal of Samples that failed QC thresholds ................................................ 24 2.6.4 Further Removal of Markers that failed QC thresholds .................................... 25 2.7 Genetic Linkage Analysis by NPL Method .............................................................. 25 2.8 Genetic Association Analysis by Family-Based Method ......................................... 28 2.8.1 Selection of Subject and Marker Data .............................................................. 28 2.8.2 Testing for Single Marker Association ............................................................. 29 2.8.3 Testing for Multi-marker Association .............................................................. 30 2.9 Statement of Contributions ....................................................................................... 30  iv  CHAPTER 3 GENOME-WIDE NON-PARAMETRIC LINKAGE ANALYSIS ................................... 40 3.1 Introduction ............................................................................................................... 40 3.2 Methods..................................................................................................................... 40 3.3 Results ....................................................................................................................... 41 3.4 Discussion ................................................................................................................. 42 3.4.1 Novel Linkages ................................................................................................. 42 3.4.2 Replication of Linkage Regions........................................................................ 43 3.4.3 Failure to replicate linkage signals ................................................................... 46 3.5 Subsequent Analyses ................................................................................................ 49 CHAPTER 4 FAMILY-BASED ASSOCIATION ANALYSIS .......................................................... 62 4.1 Testing for Association at 16p12 .............................................................................. 63 4.1.1 Introduction ....................................................................................................... 63 4.1.2 Method .............................................................................................................. 65 4.1.3 Results ............................................................................................................... 65 4.1.4 Discussion ......................................................................................................... 65 4.2 Testing for Association at 2p21 (DYX3) .................................................................. 70 4.2.1 Introduction ....................................................................................................... 70 4.2.2 Methods and Results ......................................................................................... 71 4.2.3 Discussion ......................................................................................................... 72 4.3 Testing for Association at 4q12-q13.1 ...................................................................... 74 4.3.1 Introduction ....................................................................................................... 74 4.3.2 Methods and Results ......................................................................................... 75 4.3.3 Discussion ......................................................................................................... 76 4.4 Testing for Association at 6p22.2 (DYX2) ............................................................... 78 4.4.1 Introduction ....................................................................................................... 78 4.4.2 Methods............................................................................................................. 79 4.4.3 Results ............................................................................................................... 79 4.4.4 Discussion ......................................................................................................... 80 4.5 Correcting for multiple comparisons ........................................................................ 82 CHAPTER 5 CONCLUSION ...................................................................................................... 92 REFERENCES .................................................................................................................................. 95 APPENDIX1 .................................................................................................................................. 119 APPENDIX2 .................................................................................................................................. 121 APPENDIX3 .................................................................................................................................. 123 APPENDIX4 .................................................................................................................................. 126   v  LIST OF TABLES Table 1 - Summary of published linkage and association findings for dyslexia. ......................... 14  Table 2 - Structures of 100 dyslexia families ............................................................................... 36  Table 3 - Summary statistics for the sets of dyslexia families used in the genetic analyses. ....... 37  Table 4 - Ten samples that failed QC measures ........................................................................... 38  Table 5 - Number of genotyped SNP markers found in the four regions that were tested for association to PCD. ............................................................................................................... 39  Table 6 - Results from single marker association and Haplotype TDT association for 16p12. ... 83  Table 7 - (7A) Genes found in region chr16:17,161,049-23,871,593 bp. (7B) Copy Number Polymorphisms (CNP) found in region chr16: 17,161,049-23,871,593 bp. ......................... 84  Table 8 - Results from single marker association and Haplotype TDT association for 2p21. ..... 87  Table 9 - Genes found in region chr2:43,105,656-45,947,264 bp. ............................................... 88  Table 10 - Results from single marker association and Haplotype TDT association in region 4q13-4q12. ............................................................................................................................ 89  Table 11 - (11A) Genes found in region chr4:57,291,670-61,950,748 bp. (11B) Copy number polymorphisms (CNP) found in region chr4:57,291,670-61,950,748 bp. ............................ 90  Table 12 - Results from single marker association and Haplotype TDT association for 6p22.2. 91  Table 13 - 2-SNP haplotype TDT p-values for all haplotypes in a window when one of the two markers was associated with PCD in the single marker association tests for the 16p12 region. ................................................................................................................................. 119  Table 14 - 2-SNP haplotype TDT p-values for all haplotypes in a window when one of the two markers was associated with PCD in the single marker association tests for the 2p21 region. ............................................................................................................................................. 121  Table 15 - 2-SNP haplotype TDT p-values for all haplotypes in a window when one of the two markers was associated with PCD in the single marker association tests for the 4q12 region. ............................................................................................................................................. 123  Table 16 - 2-SNP haplotype TDT p-values for the rs16888748|rs3804320 haplotypes in the 6p22.2 region. ..................................................................................................................... 126  vi  LIST OF FIGURES Figure 1:  Sample of one of the four extended pedigrees that were split in half, before and after splitting of families to be ≤22bits ......................................................................................... 31  Figure 2: Large extended pedigree with 31 individuals, before and after splitting of families to be ≤22bits ................................................................................................................................... 32  Figure 3: Large extended pedigree with 99 individuals, before and after splitting of families to be ≤22bits. .................................................................................................................................. 33  Figure 4: Summary statistics for set of families used in NPL analysis. ....................................... 35  Figure 5: Results for non-parametric linkage analysis for chromosomes 1-22 and the X chromosome. ......................................................................................................................... 50    vii  LIST OF ABBREVIATIONS ADHD  attention-deficit hyperactivity disorder AFBAC affected family-based controls ASP  affected sibpair BRLMM Bayesian Robust Linear Model with Mahalanobis distance classifier    (genotype calling algorithm) CDCV  common disease common variant (hypothesis) cM  centiMorgan CNP  copy number polymorphism CNS  central nervous system CNV  copy number variant DM   Dynamic Model (genotype calling algorithm) FBA  family based association GABA  gamma-aminobutyric acid Gb  gigabytes GCA  general cognitive ability GCOS  GeneChip® Operating Software GTYPE GeneChip® Genotyping Analysis Software GNF  Genomics Institute of the Novartis Research Foundation HLA  human leukocyte antigen HWE  Hardy-Weinberg equilibrium IBD  identity-by-descent IBS  identity-by-state IR  information ratio LD  linkage disequilibrium LGN  lateral geniculate nucleus LOD / lod log of the odds ratio Mb  megabase MLS  maximum LOD scores MRI  magnetic resonance imaging mRNA  messenger ribonucleic acid  viii  ng  nanogram NPL   non-parametric linkage (or non-parametric lod) PCD  phonological coding dyslexia PCR   polymerase chain reaction PKC  protein kinase C QC  quality control QTL  quantitative trait loci RAM  random access memory RAN  rapid automatized naming RT-PCR reverse transcriptase polymerase chain reaction SLE  systemic lupus erythematosus SLI  specific language impairment SNPs   single nucleotide polymorphisms (also used to refer to SNP markers) SPCH  speech and language disorder SSD   speech-sound disorder STR  short tandem repeat (also used to refer to STR markers) TDT  transmission/disequilibrium test UCSC  University of California, Santa Cruz VIF  variance inflation factor χ2   Chi-squared     ix  ACKNOWLEDGEMENTS  I thank my thesis advisor, Dr. LL Field, for her good-humoured supervision and direction. I am grateful to my lab-mates: Ella, Danielle, Karey and Shao for all their help, support, and for making me laugh.  In particular, I thank Ella, Danielle and Shao for the long hours spent generating the array-data and Karey for her statistical advice and nifty skills in R. I am indebted to my family for putting up with me and all of my requests. Finally, I acknowledge my mother, without whom I could not have accomplished this endeavor, and my eighth-month old daughter Mayumi who never sleeps, but always smiles.  1  CHAPTER I INTRODUCTION / REVIEW OF THE LITERATURE 1.1  Developmental Dyslexia: Phenotype and Prevalence Dyslexia is a learning disability characterized by difficulty with language-related functions that cannot be explained by deficits in intelligence, educational opportunity, motivation or sensory acuity. Dyslexia affects 3%-10% of school-age children (Yule and Rutter 1976; Rutter 1978; Hynd and Cohen 1983; DeFries 1989; Rumsey 1992) and accounts for 80% of all learning- disability cases, making it one of the most frequently diagnosed childhood learning disorders (Shaywitz 1998; Lerner 1989). Dyslexia has socially important consequences since learning difficulties that manifest at an early age can affect the cognitive, social and emotional development of a child. In order to appreciate the specific reading problems associated with dyslexia, it is necessary to understand how the brain recognizes language in a hierarchical order. Firstly, words must be decoded at a phonological level, breaking words into separate small units of sound called phonemes, before they can be recognized at higher levels in the hierarchy, that is before the semantics (the meaning of words), syntax (grammatical structure) and connection of words to form sentences can be conceptualized (Shaywitz et al 1998). According to the phonological deficit theory, developmental dyslexia is characterized by difficulties in single word decoding due to problems in phonological processing that result from a biological abnormality of the brain (Stanovich 1988, Frith 1998, Ramus et al 2004). Phonological processing is a procedure that enables written words to be translated into spoken words through the identification of the individual phonemes that make up words and the ability to establish a connection between phonemes and graphemes, the basic units of the written language (Ramus et al 2004). Various prevalence estimates for dyslexia have been reported, depending on the population’s spoken language and the rigor of diagnosis. Dyslexia prevalence ranges from 5-17% in English speaking countries and, according to the Canadian Dyslexia Association, approximately five million Canadians have dyslexia. One reason for the wide range in prevalence estimates is the use of inconsistent diagnostic criteria for different dyslexia phenotypes. In addition, the variability in prevalence estimates is related to the orthographic complexity of the language  2  spoken in different countries. The English language has a deep orthography (a set of conventionally used written symbols that represent the sounds of a language), which consists of 44 phonemes that can be combined in over 1120 different ways. By comparison, Italian has a shallow orthography, with only 25 phonemes that can be combined in 33 ways (Helmuth 2001). It is considered easier to learn how to read in a language with shallow orthography, where the letters of the alphabet correspond more uniquely to each phoneme (i.e. a high grapheme- phoneme correspondence), than in languages with deep orthography (Paulesu et al 2001). For example, the estimated prevalence of dyslexia in Italy was reported as half of that in the United States (Lindgren et al 1985). Early studies of dyslexia noted a male predominance, suggesting that males may be more predisposed to this learning disorder.  This male excess, however, is less evident in family and population samples than in clinical samples and could be the result of a referral bias (Wolff and Melngailis 1994; Shaywitz et al. 1990; Vogel 1990; DeFries 1989). Recent studies indicate that while gender-specific factors likely exist, they are not as strong as originally supposed (Pennington et al. 1991, Rutter et al. 2004, Liederman et al. 2005). Sex differences in normal phonological processing (Shaywitz et al.1995) and/or the presence of X-linked recessive susceptibility genes are plausible biological explanations for a male predominance in dyslexia. Many studies have reported comorbidity between attention-deficit hyperactivity disorder (ADHD) and dyslexia (August and Garfinkel 1989, 1990; Gilger et al. 1992; Purvis and Tannock 1997; Willcutt et al. 2000, 2001, 2005; Gayan et al. 2005). As many as 20-40% of children with dyslexia also present with ADHD (Dykman and Ackerman 1991, Willcutt et al 2000). Twin studies suggest that the two disorders share some aetiological risk factors and that genes with pleiotropic effects (where one gene affects multiple phenotypic traits) are involved (Wilcutt et al 2007). It is established that dyslexia is a neurodevelopmental disorder that obstructs language acquisition and processing (Chase et al. 1996; Habib 2000). However, the study of dyslexia has been obscured by the considerable phenotypic variability among individuals classified as having dyslexia and the lack of accord in diagnosis criteria. When assigning a qualitative dyslexia phenotype, individuals are commonly classified as affected if test scores reveal a reading ability  3  that is two years below that expected by chronological age despite normal verbal IQ (Williams and O’Donovan 2006). Specialists generally agree that the most common form of dyslexia manifests as a problem with processing the basic phoneme units of language (Shaywitz 1996; Van Orden and Goldinger 1996; Manis et al. 1997; Snowling 2000). However, studies have employed numerous psychometric tests to measure the different components of reading. Age- standardized test scores for each of these reading components have been used to derive quantitative dyslexia phenotypes. Attempts to perform a meta-analysis of results from different studies (Grigorenko 2005; Rochell and Talcott 2006), though, have been complicated by the absence of universally accepted thresholds for the diagnosis of developmental dyslexia. 1.2 Theories of Developmental Dyslexia The ability to read depends on the integration of numerous complex cognitive processes, including: phonological awareness, phonological (de)coding, verbal short-term memory, orthographic coding, rapid automatized naming (RAN, a measure of processing speed) and auditory and visual processing (or the tactile sense for individuals who read Braille). Phonological awareness, the ability to recognize and consciously manipulate speech sounds, such as phonemes and syllables, can be tested by rhyming, syllable counting, and phoneme deletion tasks (Temple et al. 2003, Ramus 2004). Phonological coding, the ability to string together phonemes to read aloud written words, is generally tested through the reading of pseudowords. Verbal short-term memory, which enables known words to be recalled directly from memory instead of being dissembled into phonemes, can be assessed with non-word repetition tests (Ramus 2004). The phonological model of dyslexia argues that people with dyslexia have impaired reading ability because they have a deficit in basic phonological processing: an impaired ability to deconstruct written words into phonemes prevents word identification and activation of intact higher level linguistic processing. The phonological model of dyslexia is supported by the observation that people with dyslexia have difficulty reading yet remain intellectually capable of processing complex thoughts and ideas. Orthographic skill problems, involving the visual appearance of words, are present in a smaller subset of dyslexic patients (Olson et al 1989). An orthographic coding task involves  4  irregular word reading (e.g. yacht), while an orthographic choice test involves choosing the right word (e.g. between ‘rain’ and ‘rane’). It has been suggested that poor orthographical letter processing, also known as poor grapheme-phoneme mapping, is as a component of the phonological deficit and a consequence of poor phonological awareness (Ramus 2004). However, other researchers believe orthographic processing is independent from phonological processing and that this skill is distinctly impaired in individuals affected with dyslexia (Castles and Coltheart 1993). The magnocellular theory hypothesizes that dyslexia results from abnormalities in the magnocellular layers of the lateral geniculate nucleus (LGN) of the thalamus (Livingstone et al 1991; Stein 1993; Stein and Walsh 1997). These abnormalities may be responsible for the impaired perception of moving stimuli, which has been repeatedly observed in a portion of dyslexic children (Lovegrove et al. 1980, 1986; Breitmeyer 1993; Lehmkuhle et al. 1993; Cornelissen et al. 1995; Kubova et al. 1996; Demb et al. 1998; Witton et al. 1998; Schulte-Körne et al 2004). Neurons in the magnocellular (transient) visual pathway respond to rapidly changing, gross detail stimuli (that is, stimuli of low spatial frequency and low contrast) and are involved in the control of eye movements (Eden et al. 1996). Conversely, the parvocellular (sustained) pathway appears normal in dyslexia. There is also evidence that dyslexics are impaired at auditory tasks requiring rapid processing (Tallal and Piercy 1975; Tallal 1980; Tallal and Stark 1982; Tallal et al. 1993; McAnally and Stein 1996; Witton et al. 1998; Fitch & Tallal 2003), suggesting magnocellular-like defects in the auditory system (Galaburda et al.1994) and lending support to a rapid auditory temporal processing theory. A direct correlation between visual and auditory deficits and the degree of phonological deficit has also been noted (Demb et al 1998; Witton et al 1998).  These lines of evidence have led to a general temporal-processing-deficit theory of dyslexia (Stein and Walsh 1997, Stein 2001), where the observed sensory deficits in dyslexic subjects are attributed to dysfunction in the magnocellular neuronal cell lines involved in processing time for all sensory systems. Accordingly, the inability to perceive rapid elements of speech results in phonological deficits that hinder the multifaceted process of reading. However, Paul et al. (2006b) presented evidence against a physiologic auditory temporal deficit in dyslexia and suggested the phonological processing defect occurs at a higher level.  5  Also worth mentioning is the double-deficit hypothesis and the cerebellar deficit theory. The double-deficit hypothesis postulates that deficits in both phonological processing and the speed of processing (RAN) are responsible for the reading impairments seen in dyslexia (Wolf and Bowers 1999). Rapid naming tasks test lexical retrieval, revealing how quickly the subject can retrieve a word’s phonemic constituents from her short-term memory. Individuals affected with dyslexia have shown impairments in either rapid naming or phonological processing, while deficits in both skills have been reported for a group of more severely affected dyslexic individuals (Wolfe 1999). Therefore, many researchers consider these two as separate deficits in dyslexia (Snowling 1991). The cerebellar deficit theory associates developmental dyslexia with an impairment of the cerebellum that disrupts the performance of automatic skills, including motor control and other automatized cognitive processes (Nicolson et al 2001). This theory is supported by observations of poor balance, coordination, and time estimation in individuals with dyslexia (Fawcett et al 1996; Fawcett and Nicolson 1999). Many symptoms associate with developmental dyslexia and different affected individuals can have distinct deficits; therefore, it is impractical to expect any one theory to explain the etiology of reading disability. Dyslexia is likely a heterogeneous disorder with different underlying deficits that can act alone or in combination in diverse cases. The various theories of developmental dyslexia need not be mutually exclusive. 1.3 Neurological Studies of Dyslexia The human brain consists of a left and right hemisphere and specialization of the left hemisphere for language, in the parietal-temporal-occipital complex, was one of the earliest observations of brain asymmetry. Until recently, attempts to understand cognitive deficits in humans were restricted to post-mortem examinations of patients with brain injuries. Neuroanatomical studies that compared the brains of dyslexic individuals to those of fluent readers discovered anatomical and functional differences, confirming that dyslexia was indeed a neurobiological condition that affects processing information needed for normal reading, writing and spelling skills. Dyslexic brains were reported to more often show a reduction or even reversal of the asymmetry (in normal brain the left hemisphere is larger than the right  6  hemisphere) in the planum temporale, a language area of the left temporal-parietal lobe (Hier et al. 1978; Hynd et al. 1990; Duara et al. 1991; Shapleske et al. 1999; Hugdahl et al. 2003). More frequent developmental anomalies of the left cerebral cortex were reported (Galaburda et al. 1985; Heim and Kiel 2004). Ectopias, small nests of malpositioned neurons, and microgyri, a more severe abnormality in cortical layer architecture, have been discovered in post-mortem studies of dyslexic brains (Galaburda et al 1985). These areas of cortical disarray are believed to be the consequence of abnormal neuronal migration during fetal development of the neocortex. More recently, anatomical observations have been corroborated by functional neuroimaging studies that show reduced activation of the posterior left temporal lobe in dyslexic brains during word reading tasks (Rumsey et al. 1997; Shaywitz et al. 1998, Salmelin et al. 1996; Paulesu et al 2001).  A functional imaging study that tested responses to sound stimuli in dyslexic subjects has also reported reduced asymmetry in the planum temporale auditory cortex activation pattern (Paul et al. 2006a). Overall, these data suggest that dyslexic individuals have a more symmetrical planum temporale organization, which may reflect some form of disruption in development of the usual cerebral lateralization and asymmetry. Nevertheless, there is a degree of phenotypic variation between individuals and none of these anatomical features is definitive of dyslexia. 1.4 Genetics of Dyslexia 1.4.1  Heritability and Twin Studies The past 50 years have established dyslexia as a significantly heritable trait, with family and twin studies revealing a strong genetic component underlying this learning disability. Family history is considered one of the most important risk factors for dyslexia, and both the risk and severity increase with the number of parents affected (Wolff and Melngailis 1994; Shaywitz and Shaywitz 2003). The concordance in the diagnosis of developmental dyslexia is higher in monozygotic twins than in dizygotic twins (83% vs. 29%, Bakwin 1973; 68% vs. 38%, DeFries and Alarcon 1996), indicating a strong genetic basis for dyslexia. Even among dyslexia- discordant twins, a higher correlation of word recognition and spelling scores is observed in monozygotic twins than in dizygotic twins (DeFries et al. 1987). Estimates of heritability for dyslexia (proportion of total variability due to additive genetic factors) range from 40-70% (Gayan and Olson, 2003); however, Olson et al. (1989) found a particularly high heritability  7  (0.93) for the phonological coding component of word recognition in a study of twins with at least one dyslexic individual. Although twin studies clearly show a genetic basis for dyslexia, the inheritance mode is unclear and many genetic models have been proposed (Hallgren 1950; Finucci et al. 1976; Lewitter et al.1980; Pennington et al. 1991; Chapman et al. 2003). Dyslexia has revealed itself as a “complex trait,” that is, a trait that does not generally show simple Mendelian inheritance and for which genetic heterogeneity and environmental interactions are most probably involved. 1.4.2 Linkage and Linkage Disequilibrium Studies Researchers have reported significant genetic linkage (the tendency for two alleles at two loci that lie close to each other on the same chromosome to be inherited together) between a hypothesized dyslexia locus and mapped genetic markers from multiple chromosomal regions. Association studies (that test for linkage disequilibrium, the nonrandom association of alleles at different polymorphic sites) have also identified markers, located within dyslexia-linked regions, which show significant association with the dyslexia trait. These studies have employed both qualitative dichotomous definitions of dyslexia and quantitative phenotypes based on reading measures (quantitative trait loci, QTL). However, the comparison of results is complicated by inconsistent phenotypic diagnosis criteria. Nevertheless, there is now strong evidence for linkage of dyslexia to 15 specific regions. To date, nine dyslexia loci have been designated gene symbols DYX1-DYX9 by the Human Gene Nomenclature Committee. These nine dyslexia loci are located on chromosomal regions: 15q21 for DYX1 (Smith et al. 1983, 1991; Grigorenko et al. 1997; Schulte-Körne et al. 1998; Nothen et al. 1999; Morris et al. 2000; Chapman et al. 2004; Bates et al. 2007), 6p22.2 for DYX2 (Smith et al. 1991; Cardon et al. 1994, 1995; Grigorenko et al. 1997, 2000; Fisher et al. 1999, 2002a; Gayán et al. 1999; Turic et al. 2003; Francks et al. 2004; Deffenbacher et al. 2004; Cope et al. 2005), 2p15-p16 for DYX3 (Fagerheim et al. 1999; Petryshen et al. 2002; Kaminen et al., 2003; Bates et al. 2007),  6q12-q14.1 for DYX4 (Petryshen et al. 2001; Bates et al. 2007), 3p14.1-q13 for DYX5 (Nopola-Hemmi et al. 2001; Fisher et al 2002; Bates et al. 2007), 18p11 for DYX6 (Fisher et al. 2002a; Bates et al. 2007), 11p15.5 for DYX7 (Fisher et al 2002; Hsiung et al. 2004), 1p34-p36 for DYX8 (Rabin et al.1993; Grigorenko et al.,2001; Fisher et al 2002a; Tzenova et al. 2004; Bates et al. 2007), and finally  8  Xq27 for DYX9 (Fisher et al 2002a; de Kovel et al. 2004; Bates et al. 2007).  The same family set that was used to report the novel loci DYX4 (Petryshen et al. 2001) and DYX7 (Hsiung et al. 2004) and provide independent confirmation of DYX3 (Petryshen et al. 2002) and DYX8 (Tzenova et al. 2004) was used for the analyses described in this thesis. An additional six dyslexia linkages have been reported, but not yet assigned gene symbols, at 2q22.3 (Raskind et al. 2005; Bates et al. 2007), 4p15.33-p16.1 (Bates et al. 2007), 7q32 (Kaminen et al. 2003; Bates et al. 2007), 14q32 (Gayan et al. 2005), 13q22 (Fisher et al. 2002a), and 17p13.3 (Bates et al. 2007). Table 1 presents a summary of the studies that have reported significant findings of linkage and/or association for dyslexia. The linkage and association results suggest that multiple genes contribute to the genetic risk for dyslexia; in most cases, the ease of replication suggests that these genes have relatively major effects. Linkage disequilibrium mapping studies are now attempting to isolate candidate genes within these previously-identified dyslexia-linked regions. ROBO1 has been identified as a possible candidate for DYX5, as reduced expression of this axon guidance receptor gene shows association with dyslexia (Hannula-Jouppi et al. 2005).  The most replicated finding has been for chromosome 6p, where cumulative evidence suggests that DYX2 susceptibility could be conferred by one of two adjacent neuronal migration candidate genes: KIAA0319 (Francks et al. 2004, Cope et al. 2005; Paracchini et al. 2006), or DCDC2 (Deffenbacher et al. 2004, Meng et al. 2005, Schumacher et al. 2006), or both (Harold et al. 2006). However, not all studies have been able to detect linkage to this region (Field and Kaplan 1998; Schulte-Korne et al 1998; Nothen et al 1999; Petryshen et al 2000; Nopola-Hemmi et al 2001; Chapman et al 2004). It is likely that locus heterogeneity, perhaps with genes acting in a number of interacting networks, influences the manifestations of the dyslexia phenotype.  In a preliminary search for two-locus interaction effects, researchers in the Field lab have detected significant two-locus interaction effects that were also replicated in the reverse direction with p ≤ 0.01 (Bell et al. in prep). One of these interactions, between DYX8-DYX2 (1p-6p), has been previously reported (Grigorenko et al. 2001) and Bell et al. (in prep.) provide independent replication of this two-locus interaction. Many potential candidate susceptibility genes for dyslexia can be suggested. Neurotransmitters and their receptors, such as dopamine and gamma-aminobutyric acid (GABA),  9  are obvious candidates since they are integral components of neuronal development. The serotonergic system presents other possible candidates since serotonin is widely distributed throughout the brain and appears to be involved in early brain development. Previously identified candidate genes shown to be linked or associated with autism, ADHD, speech-sound disorder (SSD), specific language impairment (SLI), speech and language disorder (SPCH) and the autoimmune disorder systemic lupus erythematosus (SLE) should also be considered due to commonalities between dyslexia and these disorders. Nor should susceptibility genes for bipolar disorder be ignored: it is possible that a link between bipolar disorder and dyslexia may exist since the majority of cases of bipolar disorder also involve ADHD (Wingo and Ghaemi, 2007), and ADHD and dyslexia share a high degree of comorbidity. 1.5 Complex Trait Gene Mapping Genetic mapping is the first molecular step to isolating a gene, where the inheritance pattern of a trait is compared with the inheritance pattern of chromosomal regions. There are two alternative ways to search for the causative genes for a disorder: a genome-wide scan versus a candidate gene study. Candidate gene studies are driven by a prior hypothesis that the gene is a plausible candidate due to its biological function (functional candidate) or position within a region of the genome previously linked to the disorder (positional candidate). The genome-wide linkage scan is a hypothesis-free approach that attempts to identify genetic markers that segregate with the disease trait more often than expected by chance, suggesting that the marker locus and disease locus are genetically linked. Two main statistical methods, linkage analysis and association (linkage disequilibrium, LD) analysis, are used to map genetic traits. Linkage analysis tests for the co-segregation of a polymorphic genetic marker locus with disease status in families that have at least two transmissions (siblings) to compare and at least one affected offspring. Linkage is the tendency for alleles at two loci in close proximity on the same chromosome to be transmitted together to gametes generated during meiosis. For the linkage analysis of a Mendelian trait, the two transmissions can consist of an affected child and unaffected child; however, for the linkage analysis of a complex trait with reduced penetrance, the unaffected child is less informative and studies are usually designed to test multiple affected  10  offspring. Linkage analysis is based on the caveat that recombination (crossing-over) between two loci occurs at a rate that is directly related to the distance between them. Recombination frequencies (expressed in centiMorgans, cM) are determined between each marker locus and the hypothetical disease locus, and the overall likelihood of linkage is tested by calculating the Logarithm of the Odds (lod score), which is the ratio of the likelihood of the observed data given linkage divided by the likelihood of no linkage (i.e. free recombination). In parametric linkage analysis, a precise genetic model is specified - this includes the mode of inheritance, allele frequencies, and penetrance for each disease locus genotype. If the correct model is specified, then parametric analyses have more power to detect linkage than non-parametric methods (Abreu et al. 1999). Non-parametric analyses, however, are generally considered more robust for mapping complex genetic traits, since they do not require definition of any genetic model (Kruglyak et al 1996). This method, also known as the allele-sharing method, tests which chromosomal regions show elevated allele sharing between relatives with the same disease or trait phenotype (Fisher and DeFries 2002). The shared alleles are identical-by-descent (IBD), which means that both alleles are descended from the same ancestral allele, rather than identical- by-state (IBS), where two alleles are the same variant, but may be inherited from different ancestors.  In order to identify IBD alleles in siblings, the parental genotypes are required. The simplest version of a non-parametric linkage (NPL) analysis is the affected sibpair (ASP) method, where a number of nuclear families with multiple affected siblings are collected. However, the NPL method can be applied to large extended pedigrees as well. The resolution of linkage studies for complex traits is generally no better than 5-10 cM (encompassing, as a rough estimate, approximately 50-100 genes), therefore association (linkage disequilibrium) studies are usually employed to further refine the linked region, identify plausible candidate genes, and even identify causative variants. Linkage disequilibrium is defined as the tendency of allelic variants at two loci to be associated in the population more often than would be expected by chance. Association between a disease and alleles at a specific marker is assumed to be due to linkage disequilibrium between the disease locus alleles and the marker alleles. While it is unlikely that one of the tested marker loci actually contains the causative disease variant, the presence of linkage disequilibrium between the causative allele and tested marker alleles allows linked neutral marker alleles to show association with the trait.  11  Association between marker alleles and tightly linked disease mutations will persist for many generations, but alleles linked to older disease mutations will have had more time for recombination to break down the allelic association. In other words, linkage disequilibrium is a function of both inter-locus distance and time. Association studies test for a statistical association between marker alleles and the disease phenotype in populations, by comparing allele frequencies between cases and controls, or by using family-based association (FBA) methods in nuclear families. FBA requires a larger sample size than case-control methods to detect significant associations (Risch and Merikangas 1996), but it overcomes the impediment of population stratification that can produce false positive associations in improperly matched case-control studies. The standard FBA test is the transmission disequilibrium test (TDT, Spielman et al 1993), which identifies genetic associations by comparing the transmission of marker alleles from heterozygous parents to affected children with the 50% expected by random chance transmissions. This design is robust, since the untransmitted alleles act as controls for the transmitted alleles, but has reduced power since a portion of the parent-child trios will be uninformative when parents are homozygous at those markers. A critical assumption of association mapping is that the common disease common variant (CDCV) hypothesis is true (Cardon and Bell 2001). Crucial to this hypothesis is the principle that only a few genetic loci affect disease susceptibility and that each of these loci have a major causative variant that is prevalent in the population and resulting from an ancestral mutation. Therefore, the disease manifests only when a particular combination of these common alleles aggregate in the genome. Conversely, if the genetic risk for the disease is influenced by many rare alleles (allelic heterogeneity) at each of the multiple loci, then association mapping will have little power to identify the disease mutations. Attempts to discover the susceptibility genes for complex disorders have been confounded by the presence of many factors affecting the phenotype and heritability of the trait. Even though disease clusters among related individuals, indicating a genetic basis to susceptibility, complex disorders do not display simple Mendelian inheritance patterns. Complex disorders result from genetic variation at several loci in the genome and susceptibility can be further influenced by environmental factors, interactions between genes and the environment, and epistasis  12  (interactions among alleles at different loci). The genetic factors may display a reduced penetrance (probability of showing a disease phenotype when carrying a disease-related genotype), whether there are a few genes (oligogenic) or many genes (polygenic) contributing to disease susceptibility. The same genetic loci may be involved in disease susceptibility across different populations or different predisposing loci may be present in different populations (population heterogeneity). Another problem for the analysis of complex traits, especially for psychiatric disorders, is the existence of multiple definitions for heterogeneous disease phenotypes. The presence of phenocopies (non-genetic cases) can further obscure complex trait mapping, and genetic analyses that only examine affected individuals are especially susceptible to errors resulting from the presence of phenocopies (Lander and Schork 1994). The mapping of dyslexia susceptibility genes has been complicated by the same confounders that obscure the mapping of other complex traits.  Perhaps the most troublesome is the use of different ascertainment criteria and various phenotypic measures, which has led to the identification of numerous loci linked to differing phenotypic measures of dyslexia and reading. Furthermore, vastly different types of pedigrees are utilized: some follow no clear inheritance patterns, while others seem to exemplify Mendelian modes of inheritance (e.g. autosomal dominant). Despite these drawbacks, the genetic mapping of dyslexia-susceptibility loci has seen a higher degree of replication than usual for a complex disorder. As outlined in section 1.4.2 and Table 1, significant linkage to dyslexia has been reported in nine regions with DYX designations, eight of which have been confirmed in independent datasets, while studies identifying another six dyslexia-linked regions not yet assigned DYX symbols have also been independently replicated.  This high degree of replication suggests that dyslexia-susceptibility genes may be less numerous or have a larger predisposing effect than the susceptibility genes for many other complex traits. Both explanations bode well for efforts to isolate the functional variants.  13  1.6 Aim This study investigates the genetic susceptibility for dyslexia by searching for the major genes predisposing to phonological coding dyslexia (PCD) and locating them in the genome. A genome-wide linkage study and four targeted fine-mapping association studies are performed to screen for dyslexia susceptibility genes in 101 Canadian families with PCD. In Chapter 3, I test for linkage between SNP (single nucleotide polymorphism) markers and a PCD susceptibility locus in the families by rejecting the null hypothesis that the marker is not linked to the locus. My goal was to discover new susceptibility loci and to refine the location of previously identified dyslexia-linked regions by detecting linkage to PCD with a high-density marker set. In Chapter 4, I test a second hypothesis that regions which demonstrate linkage will show significant evidence for linkage disequilibrium (association) to markers at genes within the linked region that are reasonable candidate genes for a neurodevelopmental disorder. Alternatively, in section 4.4, I test the hypothesis that the previously identified dyslexia susceptibility locus (DYX2), which does not demonstrate linkage in our families, will also fail to show any significant association with SNP markers within this region. Both methods of analysis use SNP markers from the Affymetrix GeneChip® Human Mapping 500K Array Set; therefore, a secondary aim of this study is to determine the logistics of performing genome-wide linkage and targeted association analyses with a high-density marker set.  14  Table 1 - Summary of published linkage and association findings for dyslexia. Abbreviations: NPL= non-parametric linkage, QTL=quantitative trait loci, Trio=parents and affected child, ASPs =affected sibpairs Locus  Sample  Method  Reference  1p36‐p34 (DYX8)  English: USA (9 families)  Linkage: qualitative  Rabin et al 1993  English: USA (8 families, n=165)  Linkage: qualitative ( NPL &parametric) Grigorenko et al 2001  English: UK (89 families, 195 ASP)  Linkage: QTL  Fisher et al 2002  English: USA (119 families, 180 ASP)        English: Canada (100 families, n=914) Linkage: qualitative & QTL   Tzenova et al 2004  English: Australia (403 twin families)  Linkage: QTL  Bates et al 2007              2p15‐p16 (DYX3)  Norwegian (1 family, 36 affecteds)  Linkage: qualitative (NPL)  Fagerheim et al 1999, 2002 English: UK (89 families, 195 ASP)  Linkage: QTL  Fisher et al 2002  English: USA (119 families, 180 ASP)  Linkage: QTL     English: Canada (96 families, n=877)  Linkage: qualitative & QTL  Petryshen et al 2002  English: Australia (403 twin families)  Linkage: QTL  Bates et al 2007  2p16‐p12  English: USA (119 families)  Linkage: quantitative  Francks et al 2002     Association: quantitative     Finnish (11 families, n=97)  Linkage: qualitative (parametric & NPL) Kaminen et al 2003              2q22  English: USA (108 families, n=898)  Linkage: QTL  Raskind et al 2005  English: Australia (403 twin families)  Linkage: QTL  Bates et al 2007              3p12‐q13 (DYX5)  Finnish: (1 family, n=74)  Linkage: qualitative (parametric)  Nopola‐Hemmi et al 2001  English: UK (89 families, 195 ASP)  Linkage: QTL  Fisher et al 2002  English: USA (119 families, 180 ASP)        English: Australia (403 twin families)  Linkage: QTL  Bates et al 2007              4p16‐p15  English: Australia (403 twin families)  suggestive linkage: QTL (lod = 2.08)  Bates et al 2007   15   Table 1 - continued Locus  Sample  Method  Authors  6p22.2    (DYX2)   English: USA (19 families)  Linkage: qualitative & QTL  Smith et al 1991  English: USA (180 ASP)  Linkage: qualitative & QTL  Cardon et al 1994, 1995  English: USA (79 families, 126 ASP)  Linkage: QTL  Gayan et al 1999  English: USA (6 families)  Linkage: qualitative (parametric & NPL) Grigorenko et al 1997  English: USA (8 families)  Linkage: qualitative (NPL)  Grigorenko et al. 2000  English: UK (82 families, 181 ASP)  Linkage: QTL   Fisher et al 1999  English: UK (89 families, 195 ASP)  Linkage: QTL   Fisher et al. 2002  English: USA (119 families, 180 ASP)        English: USA (104 families, n = 392)  Linkage: QTL  Kaplan et al. 2002     Association: TDT     English: UK (101 trios)  Association: TDT  Turic et al. 2003  English: UK (77 trios)  Association: TDT     English: USA (349 families, n=1,559)  Linkage: QTL (sibpairs)  Deffenbacher et al 2004     Association: QTDT     English: USA (159 families)  Linkage:  QTL  Francks et al 2004  English: UK (264 families)  Association: QTDT     English: USA (156 families)  Association: QTDT  Meng et al 2005  English: UK (240 cases, 312 controls)  Association: qualitative (case/control)  Cope et al 2005  German (137 trios)  Association : qualitative TDT  Schumacher et al 2006  Engligh: UK (350 cases, 273 controls)  Assocaition: qualitative (case/control)  Harold et al 2006              6q11.2‐q12   (DYX4)  English: Canada (96 families)  Linkage: qualitative & QTL  Petryshen et al 2001  English: Australia (403 twin families)  Linkage: QTL  Bates et al 2007              7q32  Finnish (11 families, n=97)  Linkage: qualitative  Kaminen et al 2003  English: Australia (403 twin families)  Linkage: QTL  Bates et al 2007   16  Table 1 - continued Locus  Sample  Method  Authors  11p15.5 (DYX7)  English: Canada (100 families, n=914) Linkage: parametric  (dominant model)  Hsiung et al 2004     Association: family‐based     English: UK (89 families, 195 ASP)  Linkage: QTL  Fisher et al 2002              15q21 (DYX1)  English: USA (9 families, n=84)  Linkage: qualitative (parametric)  Smith et al 1983  English: USA (19 families)  Linkage: qualitative & QTL  Smith et al 1991  English: USA (6 families, n=94)  Linkage: qualitative (parametric & NPL) Grigorenko et al 1997  German (7 families, n=67)  Linkage: qualitative (parametric & NPL) Schulte‐Körne et al 1998        Nöthen et al 1999  English: UK (178 trios)  Association: TDT  Morris et al 2000  Finnish (55 cases, 113 controls)  Association: case‐control  Taipale et al 2003  English: Canada (148 trios)  Association:  TDT  Wigg et al 2004  English: USA (90 families, n=611)  Linkage: qualitative (parametric)  Chapman et al 2004  Italian (121 trios)  Association: TDT  Marino et al 2004  English: Australia (403 twin families)  Linkage: QTL   Bates et al 2007          17p13.3  English: Australia (403 twin families)  suggestive linkage: QTL  (lod = 1.99)  Bates et al 2007              18p11 (DYX6)  English: UK (89 families, 195 ASP)  Linkage: QTL  Fisher et al 2002  English: USA (119 families, 180 ASP)        English: UK (84 families, 143 ASP)        18p21  English: Australia (403 twin families)  Linkage: QTL  Bates et al 2007              Xq27  (DYX9)  English: UK (89 families, 195 ASP)  Linkage: QTL  Fisher et al 2002  English: USA (119 families, 180 ASP)        Netherlands (1 family n=29)   Linkage: qualitative  de Kovel et al 2004  English: Australia (403 twin families)  Linkage: QTL  Bates et al 2007   17  CHAPTER 2 METHODS 2.1 Study Design From 1992-1997, Dr. Leigh Field and Dr. Bonnie Kaplan (Departments of Medical Genetics and Psychology, University of Calgary) collected DNA and performed psychometric testing for a family-based study that aimed to identify the genetic susceptibility loci for dyslexia. The specific phenotype of impaired phonological coding skills was chosen as the basis for a genetic linkage study of dyslexia.  Impairment of phonological coding skills is a highly heritable deficit present in most dyslexia cases (Olson et al 1989); therefore, the susceptibility genes are expected to produce a phenotype with high penetrance and high likelihood of detection by genetic mapping. 2.2 Phenotype Phonological coding skills were the basis for assigning a qualitative PCD status (affected/unaffected/uncertain). Phonological coding was tested with "word attack" subtests from the Woodcock Reading Mastery Test (Woodcock 1987) and the Woodcock-Johnson Psychoeducational Test-Revised (Woodcock and Johnson 1989). Further psychometric tests were used in order to assist in determining the degree of certainty (not severity) for the assigned affection status (i.e., definite versus probable). The Auditory Analysis Test (Rosner and Simon 1971) was used to assess phonological awareness. Spelling was assessed with the Wide Range Achievement Test (Jastak and Wilkinson 1984) and general intelligence was measured with short forms of the Wechsler Intelligence Scale (Wechsler 1974, Wechsler 1981). Children (>8 years old) were diagnosed as affected if there was a ≥ 2 year difference between chronological age and reading test performance (for further details see Field and Kaplan 1998). For adults, the cutoff scores were not used as rigidly for the phenotype definition since one of the word attack subtests had published norms only through 18 years of age. Therefore, a structured interview, to collect a reported history of reading problems, was also used to assist with the certainty rating for adult participants. Rapid Automated Naming (RAN) speed, the ability to recall and name rapidly- presented visual symbols, was assessed with the Rapid Automatized Naming of Numbers Test (Denckla and Rudel 1976), but it was not used in assigning PCD status or certainty. Both Dr. Kaplan and a reading specialist independently reviewed each subject’s psychometric scores and  18  interview data and assigned subjects to one of five categories: 1 = definitely unaffected, 2 = probably unaffected, 3 = uncertain, 4 = probably affected and 5 = definitely affected. This coding scheme displayed high inter-rater reliability (kappa = 0.84), with 100% agreement for affected (1 or 2) versus unaffected status (4 or 5). Any untested individual, whether deceased or unavailable, was coded as 3=uncertain, regardless of family history. The certainty rating was used to prioritize subjects for array-genotyping (as described in section 2.3.1), to identify less- informative individuals who could be excluded from the linkage analysis (see section 2.3.2), and to select simplex trios with the most confident PCD affected status for association analyses (see section 2.8.1). If we assume that “definitely affected” subjects tend to be more severely affected, the use of more severely affected individuals who have a stronger genetic component to the dyslexia phenotype should increase the power of the NPL analysis to detect linkage (Leal and Ott, 2000). 2.3 Subjects One hundred Canadian families were ascertained on the basis of a pair of siblings possessing the phonological coding dyslexia (PCD) phenotype as a strategy to reduce the chance occurrence of phenocopies and genetic heterogeneity between families. Families with at least one dyslexic sibpair were ascertained from schools for learning disabled children in Alberta and British Columbia.  Entry criteria required the presence of a nuclear family that included at least two affected siblings, classified as probable or definite PCD by psychometric testing, as well as both parents available for testing. The final set of 100 families included 50 nuclear families and 50 extended pedigrees; each extended pedigree contained at least one “core” nuclear family with an affected sibpair (ASP) plus additional branches containing affected individuals.  All subjects were over 8 years of age and gave informed consent; however, deceased family members or those who declined participation were included, and designated as having an uncertain phenotype, if their presence was necessary to complete a pedigree structure. Family structures and affectation statistics are summarized in Table 2. A total of 919 individuals from the 100 families provided DNA samples, 554 adults and 365 children (<18 years old) at the time of assessment. 478 individuals were classified as probably or definitely affected, 294 as probably or definitely unaffected, and 147 as uncertain. The male:female sex ratio for affecteds was 1.7:1.  19  Two nuclear families had a parent of Asian-ancestry, one Chinese descent father and one Japanese descent father, while the rest of the families were of European descent.  Many families exhibited apparent autosomal dominant inheritance, and 22 families displayed a history of dyslexia on both maternal and paternal sides. 2.3.1 Selection of Subjects for Array-Genotyping 750 family members were selected for genotyping on the Affymetrix GeneChip® Human Mapping 500K Array Set. Cost limited the number of arrays that could be purchased, so siblings with an uncertain dyslexia phenotype were excluded, as well as some uncertain non-parent relatives in the extended families. This subset of individuals included 744 family members from the 100 Canadian families detailed above and six new individuals from Dr. Field’s own family, containing 3 severely dyslexic members. Therefore, 101 families were used in this current gene- mapping study of PCD. Six samples, from the 750 selected individuals, were excluded due to low genotyping call rates (see section 2.5). The final family set consisted of 856 individuals, 744 (87%) genotyped and 112 (13%) not genotyped but included for pedigree structure and phenotype information. Among these individuals, 426 were affected with PCD, 247 were unaffected and 183 were uncertain. There were almost twice as many non-founders (548) as founders (308) and slightly more males (468) than females (388). Founders are defined as individuals who do not have any parents identified in the pedigree, while non-founders are persons descended from founders. After quality control (QC) assessment (see sections 2.6.2 and 2.6.3 for further explanation), eight genotyped individuals were removed from the family set and another two individuals’ genotypes were discarded, although they were included in the family set for phenotype and relatedness information. Table 3 (columns A and B) details the summary statistics for the family sets before and after QC assessments were applied. 2.3.2 Selection of Subjects for Linkage Analysis The affected sibpairs in five families were determined to be monozygotic twins; the twin with the lower genotyping rate for the Affymetrix markers was removed from all gene-mapping analyses. In two of these families there was no other sibling present so these two sets of trios (mother, father, and affected child) were removed from the linkage analysis because trios are  20  only informative for association analysis, not for linkage analysis. It was necessary to remove twelve more individuals from the analysis and split six of the extended pedigrees into smaller families in order to perform a linkage analysis using the MERLIN program package (available at http://www.sph.umich.edu/csg/abecasis/Merlin/index.html; Abecais et al 2002). Multipoint linkage analysis is limited by the pedigree size: the default maximum family bit-size is 24 bits in MERLIN, where the number of bits = 2(#non-founders) – (#founders). A maximum of 22 bits per family was used to run the analysis on our in-house computers. The six large pedigrees were split into nineteen families: four pedigrees were split in half, the fifth had 31 individuals and was split into three families, and the sixth pedigree, with 99 individuals, was split into eight separate families (see Figures 1-3). Nine individuals were duplicated during the splitting of families so that they occur in two subfamilies; however, the transmission of alleles from parent to child was not duplicated. That is, the duplicated individual acted as a parent in one family and a child in the other family. Before the families were split, the family set consisted of 830 individuals from 99 families. Columns C and D of Table 3 show the summary statistics before and after the nine individuals were duplicated. The final family set used for the nonparametric linkage (NPL) analysis using Merlin had 839 individuals from 112 families. The average number of generations was 2.44: the minimum was two generations and the maximum was four. The average family size was 7.49 individuals, varying from four to twenty-three family members, as shown in Figure 4A.  726 (87%) individuals were genotyped, and 113 (13%) were genotype unknown. Among these individuals, 419 (50%) individuals were classified as affected, 242 (29%) as unaffected and 178 (21%) as uncertain. The subjects were 54% male and 46% female, with a male:female sex ratio for affecteds of 1.79:1 (see Figure 4B for affection status statistics, distributed by sex). 2.3.3 Selection of Subjects for Association Analyses A total of 83 simplex trios, consisting of both parents and an affected child, were used for the association analyses. These 83 trios were selected from the 101 array-genotyped families by choosing the eldest affected child who had both parents genotyped and a “definitely affected” phenotype (affection code of 5, as described in section 2.2). When there was no child that fulfilled these requirements, then the eldest child with an affection code of 4 (“probably affected”) was selected. Twelve of the 50 nuclear families had only one parent genotyped, while  21  in eight other families only the affected pair of siblings (not parents) were genotyped. These twenty families, therefore, failed the selection criteria and were excluded from the association analyses. The χ2 statistic used to test for association assumes that the data are independent observations and is not a valid test of association if more than one trio per family is used because the marker-disease locus phases during the meioses are no longer independent (Spielman et al 1996). Therefore, only one trio was chosen from each pedigree, except in the case of two extended pedigrees that were bilineal, with disease genes entering via unrelated pedigree members, allowing the selection of two unrelated trios from each pedigree. The final set of 83 simplex trios included 249 genotyped individuals: 128 (51%) were affected, 82 (33%) were unaffected and 39 (16%) were uncertain. Again, the set included slightly more males (140) than females (109), as shown in Table 3, column E. 2.4 Storage of Pedigree and Phenotype data All pedigree structures and dyslexia status data were imported into a Progeny Lab 7 database (Progeny Software, LLC) from a previous MS-ACCESS database. Progeny Lab is a commercial lab management software package that manages and combines phenotypic data with SNP and STR (short tandem repeat, including microsatellite) genotype data for whole genome or targeted association or linkage studies (http://www.progenygenetics.com/lab/index.html). Data integrity was assured by manual confirmation of all family structures and dyslexia phenotype information. Progeny Lab 7 was used to create the specific family sets required for the genetic analyses (as described in sections 2.3.2 and 2.3.3) and to output the data in file formats suitable for the genetic analysis programs. 2.5 Array-Genotyping for SNP Markers For each subject, DNA was extracted from a 14ml whole blood sample, according to a standard salting out procedure (Miller et al. 1988). DNA samples were genotyped for 262,217 SNPs using the Nsp I array from the Affymetrix GeneChip® Human Mapping 500K Array Set. This set comprises two 250K arrays, each using a different restriction enzyme: Nsp I or Sty I. A 250ng sample of genomic DNA from each subject was processed for hybridization to the Nsp I  22  array in 96-well PCR (polymerase chain reaction) plates according to the manufacturer’s protocol (https://www.affymetrix.com/support/downloads/manuals/500k_assay_manual.pdf) by one of three technicians: Elzbieta Swiergala, Danielle Nguyen and Shao Lu. Arrays were hybridized in a GeneChip® Hybridization Oven 640, washed on a GeneChip® Fluidics Station 450 and scanned with a GeneChip® Scanner 3000, according to the manufacturer’s protocol (Affymetrix, Inc), by the same three technicians. The scanner and fluidics station were controlled by GeneChip® Operating Software (GCOS) 1.4 and the initial SNP genotype calls were made by GeneChip® Genotyping Analysis Software (GTYPE) 4.0 using the Dynamic Model (DM) algorithm (Di et al. 2005). Thirty-seven samples had genotyping call rates (% of SNPs that were assigned a genotype) below 90% and the assays were repeated for these samples. A total of 749 samples passed a call rate cutoff of 85% (one sample was below the cutoff and discarded) using the DM algorithm. Due to limiting computing power, these 749 samples were divided into two batches of 374 and 375 samples for BRLMM algorithm calling. For each batch, the final SNP genotypes were called with the Bayesian Robust Linear Model with Mahalanobis distance classifier, using the Affymetrix® BRLMM Analysis Tool software (http://www.affymetrix.com/support/technical/whitepapers/brlmm_whitepaper.pdf). The BRLMM calling algorithm was introduced by Affymetrix during the genotyping phase of our study and is considered to produce more accurate calls than the DM algorithm. A total of 741 samples passed a BRLMM call rate cutoff of 90%, with an average call rate of 97.78%, and 722 had a call rate above 93%.  Of the eight samples with BRLMM call rates below 90%, three subjects were integral to the family-based analyses and their samples (with call rates of 87.54%, 89.36% and 89.69%) were kept, while the other five samples were dropped from analyses. The final set of genotypes from 744 subjects were stored and output in file formats suitable for analysis using the Progeny Lab 7 database (Progeny Software, LLC). 2.6 Quality Control Measures Phenotype and genotype data for 856 individuals (744 genotyped) was exported from Progeny Lab 7 into PLINK for quality control assessment. PLINK is a whole-genome genetic analysis tool set designed to generate summary statistics and to perform advanced error checking and genetic analyses for high-density marker data (http://pngu.mgh.harvard.edu/~purcell/plink/).  23  One hundred and twelve individuals were not genotyped but did have phenotype data and were necessary for pedigree structure, while the other 744 individuals were genotyped for the 262,217 SNP marker set. PLINK accepts multigenerational pedigree data for family-based tests, breaking them down into nuclear family units to perform analyses (Purcell et al 2007). Summary statistics were generated for missing genotype rates (per individual and per SNP), minor marker allele frequencies, Hardy-Weinberg equilibrium (HWE) test failure, single-SNP non-Mendelian transmission rates (per pedigree and per SNP), and individual sample heterozygosity rates. These statistics were then used to filter out SNP markers and individuals that did not pass inclusion thresholds. 2.6.1 Initial Removal of Markers that Failed QC Thresholds Stringent inclusion thresholds were applied that significantly reduced the number of SNPs in the final marker set, since undetected genotyping error can encumber linkage and association studies. 57,275 SNPs had a minor allele frequency (MAF) less than 5% in the founder individuals, consisting of 192 genotyped persons without parents in the pedigree. 85,717 SNPs were missing genotypes in more than 2% of the 744 genotyped samples. SNPs that failed either of these thresholds were removed, reducing the number of SNPs in the marker set to 134,051. Thus the remaining 134,051 SNPs had a MAF of at least 5% and were successfully called in at least 98% of genotyped persons. 2.6.2 Resolution of a Sample Mix-up Results from the PLINK summary statistics revealed that six samples had been swapped during the genotyping process. Five pedigrees had per-family Mendel error rates above the recommended 5% threshold, displaying an error rate between 19-23%. Genome-wide identity- by-descent (IBD) estimates were determined for all pairs of individuals in the data set, to identify potential sample swaps or duplications.  PLINK calculates genome-wide IBD sharing coefficients (the probability of sharing 0, 1, or 2 alleles IBD) and uses the   metric to measure the proportion of IBD sharing between two individuals, where =P(IBD=2)+0.5*P(IBD=1).  A  value of approximately 0.5 is expected for any parent-child or sibling comparison, while a  24  =1 indicates a sample duplication and a   just below 1 indicates a monozygotic twin pair (theoretically  should be 1, but it ends up being slightly less due to different genotyping errors in each sample). Seven samples returned  estimates far below the expected 0.5 result for IBD sharing between these samples and their recorded parents and siblings.  Instead, sample 2 displayed a ≈0.5 value with each parent and sibling from the pedigree for sample 7, and sample 7 displayed a ≈0.5 value with each parent and sibling from the pedigree for sample 2. This suggested that samples 2 and 7 had been swapped.  The same results were observed between sample 3 and sample 6, and between sample 4 and sample 5. Sample 1 did not return a value close to 0.5 with any other individual in the dataset. Laboratory records revealed that these seven samples had been hybridized sequentially onto arrays at the same time, and a reference DNA sample occupied the eighth place; therefore, it was concluded that the order had been reversed during the hybridization process. Reference DNA was hybridized onto an array, in the place of sample1 and array data for sample 1 was never genotyped. The other six sample errors were thereby resolved and, as expected, the Mendelian error rate dropped below the threshold for the corresponding families when the QC measure was reapplied. 2.6.3  Removal of Samples that failed QC thresholds A total of nine samples were removed because they failed QC assessments. One sample gave discordant results in a sex check process that compared the reported sex to the genomic sex, determined by X chromosome heterozygosity estimates, and two other samples had ambiguous results for the genomic sex. The former supposedly-male sample, which was likely DNA from a female sibling because the sample did not have an inflated Mendel error rate, was discarded. The latter two samples were discarded since they also displayed excess intra-individual genome-wide heterozygosity, (F< -0.10). The F value is an estimate of the inbreeding coefficient, based on a calculation described in Purcell et al 2007 that compares the observed versus expected number of homozygous genotypes within an individual. The F value analysis was applied to a subset of the autosomal marker data, pruned to be in approximate linkage equilibrium, as described in section 2.7. A strongly negative F value (excess heterozygosity) suggests DNA contamination or some other factor such as extreme outbreeding, for example. Interestingly, our family set included two  25  nuclear pedigrees having one parent of European descent and the other of Asian descent; as expected, the offspring showed increased intra-individual heterozygosity (-0.08<F<-0.06). These two families remained included in the analyses since this decrease in F was expected and not due to DNA contamination. A fourth sample, the father of a nuclear family that did not display the expected father-child   value and created a 20% per-family Mendel error rate, was also discarded. Finally, IBD pair-wise estimation identified five sets of monozygotic twins; the twin with the lower genotyping rate was removed from all subsequent analyses. All ten problematic samples, listed in Table 4, were discarded (including sample 1 that was never genotyped due to the mix up with reference DNA, see section 2.6.2); however, two individuals were necessary for pedigree structure and were included in the analyses, but with all marker genotypes cleared. 2.6.4 Further Removal of Markers that failed QC thresholds Once these ten problematic samples were identified, new PLINK export files were generated from Progeny Lab 7 and all summary statistics were applied to the new sample set.  The new sample set consisted of 134,051 SNP markers and 848 individuals: 734 individuals had genotyped data and 114 did not. As expected, all individuals passed the summary statistics described above and had an average per individual genotyping rate of 99.3%. All SNPs were genotyped in more than 98% of the 734 genotyped samples and had a Mendel error rate of approximately 1%, far below the 10% default threshold in PLINK.  Since there were ten less genotyped samples than in the previous sample set, the minor allele frequencies had changed and 665 markers failed the MAF 5% frequency threshold, reducing the marker set to 133,386 SNPs. The unaffected founder population of 116 individuals was used to test HWE by the HWE exact test, which is considered more accurate for rare genotypes (Wigginton et al 2005a). Another 221 markers were not in HWE (p-value < 0.001), leaving 133,165 for further analysis. 2.7 Genetic Linkage Analysis by NPL Method   A non-parametric multipoint linkage analysis was performed using MERLIN, one of the fastest pedigree analysis computer programs available (http://www.sph.umich.edu/csg/abecasis/Merlin/index.html). Non-parametric linkage (NPL)  26  statistics, which measure sharing between affected relatives, tend quickly to normal distributions and are usually quite robust when based on a large number of pedigrees of similar size (Lander & Kruglyak 1995). The p value measures the significance of an observed NPL lod score and is determined by comparing the observed NPL lod score to the distribution of NPL lod scores expected under the null hypothesis of no linkage. MERLIN, a multipoint engine for rapid likelihood inference, uses sparse inheritance trees to represent patterns of gene flow through a pedigree and can handle denser marker maps than most other linkage analysis programs (Abecasis et al 2002). The employed algorithm does not consider the improbable recombinants in its calculations (i.e. double or triple recombinants when recombination fractions are small), which improves speed and supplies more accurate solutions in a dense marker map (Abecasis et al 2002). Genotyping errors can weaken a true linkage signal by leading to incorrect inferences about the pattern of genetic inheritance in pedigrees. However, not all SNP genotyping errors will be detectable as Mendelian inconsistencies. MERLIN’s error detection analysis identifies genotypes which suggest that an unlikely recombination event occurred, one contradicted by the genotypes at neighboring markers (Abecasis et al 2002). MERLIN’s error detection considers all data simultaneously, so the improved accuracy seen in the analysis of larger pedigrees is accompanied by a sharp increase in the computing demand for available RAM (random access memory) in order to run the analysis. The complexity of our pedigrees necessitated the use of a swap file to reduce MERLIN’s memory usage demand and prevent the program from encountering a fatal error. This swap option severely reduced the program’s speed and greatly increased the time required for MERLIN to run error detection and NPL analyses. Chromosomes 1-22 and the X chromosome were run on three Dell laboratory computers: a Linux workstation with 4 Gigabytes (Gb) of RAM and two Windows workstations with 2 Gb of RAM each. The final sample for linkage analysis contained 839 subjects from 112 pedigrees (see sections 2.2.2 and Table 3, column D). Pedigree statistics were summarized and verified through PEDSTATS, a program developed by Wigginton JE and Abecasis GR (2005). A qualitative PCD phenotype was used in the linkage analyses, including genotyped individuals with an uncertain  27  phenotype, which allowed them to contribute marker information while remaining neutral with respect to affection status. In order to reduce the linkage disequilibrium (LD) between markers, since the presence of markers in tight LD can actually inflate a linkage signal and bias classical nonparametric multipoint analyses if founders are not genotyped (Goode et al 2005), a subset of the 133,165 SNPs was generated to cover the autosomal chromosomes. The PLINK program was employed to generate a subset of 52,082 autosomal SNPs that were in approximate linkage equilibrium with each other, eliminating 77,802 SNPs found to be in LD with another marker. This LD pruning was done using the variance inflation factor (VIF) method, implemented in PLINK, which recursively removes markers within a sliding window size of 50 SNPs, shifted by 5 SNPs at each step, with a VIF threshold value of 2. As described by Purcell et al, 2007: “The VIF is 1/(1-R2) where R2 is the multiple correlation coefficient for a SNP being regressed on all other SNPs simultaneously.” After pruning for LD, the marker set was still quite dense: the median inter-marker genetic distance was 0.038 cM, 25% of the markers were within 0.014 cM of another marker, and 75% of the markers were within 0.850 cM of another marker. There were some regions of the genome where marker coverage was sparser: the maximum inter-marker distance was 4.381 cM, seen on chromosome 9, and the next largest inter-marker distance was 2.287 cM, on chromosome 17. The average maximum inter-marker distance per chromosome was 1.483 cM.  It is a common assumption that recombination rates and LD are uniformly distributed across the genome and that equally spaced SNPs will capture most of the genetic information. However, in reality the pattern of LD across a small region of the genome is unpredictable and closely linked SNPs are not necessarily in high LD. The MINX program (MERLIN for the X chromosome) was used to test for NPL between PCD and 3,281 markers on the X chromosome. These markers were not pruned for LD because the PLINK LD-based SNP pruning tool automatically removed all X chromosome markers from the marker set.  Furthermore, MINX is unpublished and there was no documentation or support for its use, nor was its performance guaranteed. Therefore, the linkage data for the X chromosome from the NPL analysis may be less accurate than the linkage data for the autosomes. The lod scores from the NPL analysis calculated by MERLIN were imported into the  28  R statistical package to generate per-chromosome graphical plots. These graphs are exhibited in Chapter 3, Figure 5. 2.8 Genetic Association Analysis by Family-Based Method 2.8.1 Selection of Subject and Marker Data Detection of significant association (LD) between SNP markers and dyslexia provides statistically independent confirmation of linkage between SNPs and dyslexia in that region. Detection of significant LD may also assist in fine-mapping the dyslexia susceptibility locus. Family-based association analyses were performed to confirm the best (highest) three linkage signals identified in the genome-wide NPL screen (as described in section 3.3) and to refine the localizations of putative PCD susceptibility genes. A forth region, at 6p22.2, was tested to see if our family set would show association with marker alleles within the DYX2 region, where significant linkages and associations with dyslexia have been reported by multiple groups (Table 1 provides a list of these studies and references). As described previously in section 2.3.3, the same set of 83 simplex trios (affected child and both parents) was used in all family-based association analyses. Four marker subsets from the set of 133,165 SNPs (described in section 2.6.3) were tested for association with PCD. Boundaries for the regions identified by the three NPL linkage peaks were established on either side of the maximum lod score at the two points where the lod scores dropped below a cutoff value of lod=1.0. The first region on 16p spanned 8.3 Mb, from 15.684-23.978Mb, and included 224 markers. The second region on 4q spanned 5.0Mb, from 57.177-62.070Mb, and included 355 markers. The third region on 2p spanned 4.4Mb, from 41.871-46.278Mb, and included 280 markers. The fourth region at 6p22.2 spanned 700kb with 28 markers, from 24.200-24.900Mb, and included both the VMP/DCDC2/KAAG1 and KIAA0319/TTRAP/THEM2 gene clusters that have been previously associated with dyslexia. The HAPLOVIEW program was used to visualize the LD structure and interlink the marker set with the Entrez gene map in each of the four regions tested for association (Barrett 2005; available at http://www.broad.mit.edu/mpg/haploview/index.php). The TAGGER SNP selection algorithm, within the HAPLOVIEW program, was used to identify a subset of SNPs that  29  captured the variation seen in the complete set (de Bakker et al 2005). The r2 statistic was employed to measure the degree of LD between two markers and acted as the pair-wise metric to assess tagging quality. Marker-marker LD values of r2>0.33 are generally considered useful in fine mapping studies (Langefeld and Fingerlin, 2007). If the r2 value between two SNPs was above a threshold of r2=0.8 (as suggested by Carlson et al. 2004), they were considered sufficiently correlated that only one of the two SNPs was included in the set of tagging SNPs. Table 5 presents the number of markers in each of the four analysis regions, both before and after the marker sets were pruned for LD. As shown in Table 5, in each of these four regions the marker set used for the NPL linkage analysis pruned to be in linkage equilibrium (column B) contains just under two thirds as many markers as those in the association sets (column C), which were LD-pruned to a lesser degree. 2.8.2 Testing for Single Marker Association The Affected Family-Based Controls program, developed by Drs. Field and Thomson (AFBAC; Thomson 1995) was used to test for evidence of an increased frequency of specific marker alleles transmitted from parents to affected offspring when compared to the parental marker alleles not transmitted. The Transmission Disequilibrium Test (TDT) (Spielman et al. 1993) was also used to test association on a subset of the AFBAC data, since TDT only considers transmissions from heterozygous parents to affected offspring. Because AFBAC uses data from all parents, whether homozygous or heterozygous at the marker, it has more power to detect a true association. In AFBAC, statistical significance is determined using a χ2 test in a 2x2 contingency table for a bi-allelic marker. Under the null hypothesis, the transmissions of parental alleles are independent events and the frequencies of transmitted and non-transmitted alleles are the same (for example, major and minor allele frequencies of 0.7 and 0.3 in both transmitted and non-transmitted alleles). The TDT uses the χ2 statistic to test for an equal number of transmissions of both alleles, for a bi-allelic marker, from heterozygous parents to an affected offspring under the null hypothesis of no association where both major and minor alleles have a 50% chance of transmission (Spielman & Ewens 1996). The TDT was applied through the PLINK program package implementation (Purcell et al. 2007).  30  2.8.3 Testing for Multi-marker Association Genotyped marker data for those SNP markers that displayed single-marker association were also analyzed for haplotype associations with the PCD phenotype in the same set of 83 trios. A two-SNP sliding window was employed to specify the particular haplotype tests. A haplotype- based TDT was performed through the PLINK program, using expectation-maximization phasing to determine the expected haplotype distribution for each individual (Purcell et al. 2007). 2.9 Statement of Contributions Dr. Leigh Field and Dr. Bonnie Kaplan (Departments of Medical Genetics and Psychology, University of Calgary) collected DNA for the 101 families. Dr. Bonnie Kaplan performed the psychometric testing. Technicians in the Field laboratory: Ella Swiergala, Danielle Truong and Shao Lu performed the array-genotyping procedure on the DNA samples. Scripts for the R statistical package were created by Karey Shumansky to generate the per-chromosome graphical plots (used in Figure 5) for the NPL analysis.    31  1.1)    1.2)           Figure 1:  Sample of one of the four extended pedigrees that were split in half, before and after splitting of families to be ≤22bits (in order to comply with the MERLIN program’s pedigree size requirement to run NPL analysis) 1.1) Original pedigree before split. 1.2) Two resulting families after split.  Legend:   (circle = female, square = male) = uncertain status = affected with PCD = unaffected status  32  2.1)             2.2)  A.       C.               B.            Figure 2: Large extended pedigree with 31 individuals, before and after splitting of families to be ≤22bits (in order to comply with the MERLIN program’s pedigree size requirement to run NPL analysis) 2.1) Original pedigree before split. 2.2)  A. B. and C. depict the three resulting families after the split. Legend:   (circle = female, square = male) = uncertain status = affected with PCD = unaffected status  33  Legend:   (circle = female, square = male) = uncertain status = affected with PCD = unaffected status 3.1)               3.2) A.          B.             Figure 3: Large extended pedigree with 99 individuals, before and after splitting of families to be ≤22bits. (in order to comply with the MERLIN program’s pedigree size requirement to run NPL analysis) 3.1) Original pedigree before split. 3.2)  A. B. C. D. E. F. G. and H. depict the eight resulting families after the split.  34  C.        D.      E.                  F.         G.     H.             Figure 3: continued  35  A.                        B.                   Figure 4: Summary statistics for set of families used in NPL analysis. A. Pedigree size statistics. B. Affection status statistics, distributed by sex.  36   Table 2 - Structures of 100 dyslexia families   50 NUCLEAR FAMILIES (both parents and their children)  14 families with 2 affected (probable or definite) children 20 families with 3 affected members 9  families with 4 affected members 6  families with 5 affected members 1  family with 6 affected members  50 EXTENDED KINDREDS (with a core nuclear family having 2 affected sibs)  25 kindreds with <5 affected members 19 kindreds with 5-9 affected members 6 kindreds with >9 affected members (one kindred each containing                               10, 11, 15, 20, 21, and 35 affected members)  PHENOTYPE TOTALS  919 Subjects Blood Sampled for DNA (901 also psychometrically tested)  554 adults, 39% affected 365 children (<18 yr. old), 72% affected  478 (52%) Probable or Definitely Affected (male:female ratio 1.7 to 1) 294 (32%) Probable or Definitely Unaffected 147 (16%) Uncertain (ambiguous results or refused testing)     37  Table 3 - Summary statistics for the sets of dyslexia families used in the genetic analyses. Column A: structure of family set that was array-genotyped. Column B: structure of family set that passed QC measures. Column C: structure of family set that was used for non-parametric linkage (NPL) analysis and conformed to the MERLIN analysis program requirements (before 9 individuals were duplicated). Column D: structure of family set used for NPL linkage analysis (after 9 individuals were duplicated). Column E: structure of family set used for family-based association analyses.      A  B  C  D  E    Family Set  Array‐ Genotyped  Family Set  after QC  Family Set for  NPL Analysis  Family Set  for NPL with  duplicates  Simplex Trio  Set     Individuals  856     848     830    839     249     Genotyped  744  (87%)  734  (87%) 717 (86%) 726  (87%)  249  (100%) Not Genotyped  112  (13%)  114  (13%) 113 (14%) 113  (13%)  0  (0%) Known  Phenotype  673  (79%)  665  (78%) 654 (79%) 661  (79%)  210  (84%) Affected  426  (50%)  419  (49%) 417 (50%) 419  (50%)  128  (51%) Unaffected  247  (29%)  246  (29%) 237 (29%) 242  (29%)  82  (33%) Uncertain  183  (21%)  183  (22%) 176 (21%) 178  (21%)  39  (16%) Males  468  (55%)  463  (55%) 455 (55%) 457  (54%)  140  (56%) Females  388  (45%)  385  (45%) 375 (45%) 382  (46%)  109  (44%) Founders   308  (36%)  307  (36%) 301 (36%) 311  (37%)  166  (67%) Non‐Founders  548  (64%)  541  (64%) 529 (64%) 528  (63%)  83  (33%)    38  Table 4 - Ten samples that failed QC measures *Note: sample 3934-202 erroneously consisted of reference DNA   Sample  Reason Sample Discarded   1902‐203  Discordant genomic and recorded sex  1937‐203  Ambiguous result for genomic sex  Contamination: F< ‐0.1  1945‐001  Ambiguous result for genomic sex  Contamination: F< ‐0.1  1950‐101  Father‐child  <<0.5  Per‐family Mendel error rate > 5%  3934‐202 *  Parent‐child and sibling  <<0.5  Per‐family Mendel error rate > 5%  1068‐202  Monozygotic twin:  =0.9993  1910‐202  Monozygotic twin:  =0.9998  1909‐202  Monozygotic twin:  =0.9994  1911‐204  Monozygotic twin:  =0.9996  1922‐202  Monozygotic twin:  =0.9999                  39  Table 5 - Number of genotyped SNP markers found in the four regions that were tested for association to PCD. Column A: number of markers before any test for linkage disequilibrium was applied. Column B: number of markers left after application of the more-stringent variance inflation factor (VIF) method to produce a set of markers in approximate linkage equilibrium. Column C: number of tag SNP markers used for the association analyses, generated by applying a less-stringent threshold of r2>0.8 to remove markers that were in high LD.         A  B  C        # of Markers        Before any   LD prune  After VIF   LD prune  Association Set   r2 > 0.8 Chromosome  Region (Mb)  Distance  16p  15.684‐23.978  8.3 Mb  224 121  176 4q  57.177‐62.070  5.0 Mb  355 120  224 2p  41.871‐46.278  4.4 Mb  280 112  185 6p22.2 (DYX2)  24.200‐24.900  700 kb  28 14  23    40  CHAPTER 3 GENOME-WIDE NON-PARAMETRIC LINKAGE ANALYSIS 3.1 Introduction A high-density genome-wide non-parametric (NPL) linkage scan was performed on our Canadian families having multiple members affected with phonological coding dyslexia (PCD). Our family set consisted of some pedigrees displaying apparent autosomal dominant inheritance of dyslexia, while other pedigrees displayed possible autosomal recessive or non-Mendelian inheritance patterns. Therefore, the NPL method was appropriate, since allele-sharing methods do not require a mode of inheritance to be specified. As described earlier (section 1.4), multiple dyslexia-linked regions have been identified and most of these linkage signals have also been independently replicated (confirmed) by various research groups from around the world. My goal in this portion of the study was to use SNP markers to perform the most dense genome-wide linkage screen for dyslexia susceptibility genes to date, in order to see if using a denser marker set would identify any new dyslexia-linked regions. 3.2 Methods 839 individuals in 112 Canadian families displaying inheritance of PCD were tested for NPL using the MERLIN program package; see section 2.3 for a description of how subjects were ascertained and Table 3, column D, for summary statistics of the family set.  A marker set of 52,082 autosomal SNPs was selected to test for linkage using MERLIN, as described in section 2.7. However, the final autosomal marker set consisted of 51,437 markers, since 645 markers were excluded because they shared the same cM position as another marker, despite having different base pair positions in the physical map. The MINX program (within the MERLIN package) was used to test for NPL with 3,281 SNPs on the X chromosome. MERLIN’s input map file uses genetic distance (measured in cM) to specify where the markers are located in the genome and requires each SNP marker to have a unique genetic distance. An Affymetrix NetAffx annotation file (http://www.affymetrix.com/support/technical/byproduct.affx?product=500k) for the Mapping 250K NSP enzyme provided an estimate of SNP genetic distances from the p-telomere of each chromosome. These genetic distances were estimated from the experimentally obtained deCODE  41  map, Marshfield map, and SLM1 map. The MERLIN NPL scoring function was used to assess sharing of alleles at each marker locus among affected relative pairs. Statistical evidence for linkage occurred when the allele sharing was significantly larger than expected under the null hypothesis of no linkage. The MERLIN program used the Kong and Cox function (Kong and Cox, 1997) to convert NPL scores into lod scores and calculate a p-value based on the empirical null distribution. This empirical p-value was calculated by determining the proportion of points in the null distribution that had increased allele sharing when compared to the observed statistic. 3.3 Results Figure 5 provides the total NPL lod scores summed across all families for the SNP markers along each chromosome. The three largest linkage peaks were seen on chromosomes 16, 4 and 2 with maximum lod scores of 2.263 (rs154533 at 16p12.1, 22.626Mb), 1.879 (rs1353401, rs1353402, and rs193302 at 4q13.1, 59.585-59.680Mb) and 1.831 (rs1377687 at 2p21, 44.195Mb). Chromosomes 3, 10, 11, 13, 15, 19, 20 and 21 showed no signs of linkage (i.e., no lod score >1) to PCD in the 122 families tested. In particular, linkage signals were absent for 3p12-q13 (DYX5), 11p15.5 (DYX7), and 15q21 (DYX1). No linkage signal was observed for the DYX2 locus on 6p22.2, but a lod score of 0.934, for 3 neighboring SNP markers at 90cM (80Mb), was observed for the DYX4 locus on 6q14.1. A maximum lod score of only 0.669 was obtained for the DYX6 locus at 18p11. Chromosomes 5, 8, 9, 12, 14, and 22 had some weak linkage signals, but none above a lod score of 1, and are not discussed further. The X chromosome and chromosomes 1, 2, 4, 7, 16, and 17 all displayed maximal lod scores above 1.0, as shown in Figure 5. The X chromosome had a lod of 1.698 at Xp11.1, over 80Mb away from the DYX9 locus, at Xq26-27, where no linkage was observed. Chromosome 1 had a lod of 1.244 at 1p31.1, over 60Mb from the more distal DYX8 locus, at 1p36-p34, where no linkage was observed. Aside from the aforementioned linkage peak at 2p21, chromosome 2 displayed a second linkage peak at 2q14.3, with a lod 1.366 for markers rs779976 and rs812952, providing support for the reported 2q22 dyslexia locus (Raskind et al 2005, Bates et al 2007). A lod score of 1.113 for 7q32.2 (at marker rs4731689 at 131cM, 130Mb) provided some support for the previously identified dyslexia linkage at 7q32 (Kaminen et al 2003; Bates et al 2007). Lod scores of 1.028 and 1.124 for 4p15.1 (rs1523166 at 52cM, 32Mb) and 17p13.3 (rs8065080 at 10cM,  42  3Mb), respectively, provide some support for the putative dyslexia loci at 4p16.1-15.33 and 17p13.3, recently suggested by Bates et al 2007. Chromosome 4 also displayed a third linkage peak at 4q23, with a lod score of 1.425 (rs4399995 at 104cM, 100Mb). 3.4 Discussion 3.4.1 Novel Linkages This NPL study has identified two novel linkages for dyslexia at 16p12.1 and 4q13.1. The maximum NPL lod scores of 2.263 and 1.879, respectively, do not exceed Lander and Kruglyak’s (1995) recommended threshold of lod 3.3 (p=4.9x10-5) to declare significant linkage, but the suggestive evidence of linkage is quite strong for a complex trait. The locus at 16p12.1 (lod 2.263 and p=6.2x10-4) is above the lod 1.9 (p=1.7x10-3) threshold for suggestive linkage and the 4q13.1 lod 1.879 (p=1.6x10-3) is right at the threshold (Lander and Kruglyak, 1995). This study detected linkage of dyslexia to 16p13.11-p12.1 with lod score above 1.0 from 15.684 – 23.978Mb. Chromosome 16p has not been previously linked to dyslexia; however, the 16p13-p11 region has shown linkage to ADHD, autism, and bipolar disorder (Fisher et al 2002; Smalley et al 2002, 2005; Ogdie et al 2003, 2004; Turic et al 2004; Barnby et al 2005; Cassidy et al 2007). Genome-wide scans in affected sibpairs have returned multipoint maximum lod scores (MLS) showing significant linkage to ADHD at 16p13: MLS= 4.2, p=5x10-6 (Smalley et al 2002) and MLS=3.73, nominal p=3.5x10-4, empiric p=0.012 (Ogdie et al 2003, 2004). A strong association between reading disability and ADHD has been widely reported (e.g. Hinshaw 1992; Clark et al 2002; McGee et al 2002; Willcutt et al 2001, 2005), as has support for common genetic risk factors underlying this association (Stevenson et al 1993; Chadwick et al 1999; Light et al 1995; Willcutt et al 2007). Multipoint linkage analysis in sibpairs affected with ADHD using a quantitative measure of atypical cerebral asymmetry (the absence of left hemisphere dominance for language) has also shown support for linkage (p<0.004) at 16p13 (Smalley et al 2005). Shaywitz et al (1992) and Pennington and Lefly (2001) have proposed the existence of common aetiological factors that underlie general reading ability and reading disability. In 2004, Loo et al identified loci contributing to the variation seen for normal reading performance in a sample of 223 ADHD affected sibpairs. They tested three highly correlated measures of reading  43  ability (spelling, reading recognition, and reading comprehension) using the Peabody Individual Achievement Test-Revised (PIAT-R; Markwardt 1989) and combined the scores into a READ factor score, using principal component analysis. They performed multipoint QTL mapping with ~400 microsatellite markers in a parametric genome-wide linkage analysis and found suggestive evidence for a common locus influencing reading and ADHD located on 16p12.2; a multipoint maximum lod score of 2.24 was observed for the READ factor and D16S3046 at 20.79Mb. This study detected linkage to 4q at two distinct locations (4q12-13.1 and 4q23), with lod scores above 1.0 at 57.177-62.070Mb and 96.466-100.980Mb. Chromosome 4 has not shown linkage in any previous studies of families with dyslexia, however, a region on 4p has been linked to reading and spelling ability in a normal sample of Australian twins without any learning disorder, as further described in section 3.4.2 (Bates et al 2007). The 4q12-q22 region has been linked to bipolar disorder in genome-wide scans: two point analysis identified linkage (lod=3.30) at 4q13.3 with marker D4S392 (70.56 Mb) in 395 UK and Irish ASPs (Lambert et al 2005) and NPL analysis identified linkage (NPL 2.23, p=0.013) at 4q21 for marker D4S3243 in 60 Irish ASP families (Cassidy et al 2007). 3.4.2 Replication of Linkage Regions Chromosomal regions 2p21, 2q14, 4p15.1, 7q32.2, and 17p13.3 all displayed maximal NPL lod scores above 1.0; however in the present study, a 1.44 lod criterion was enforced to confirm independent replication of a previously-reported linkage peak, according to Lander and Kruglyak’s (1995) recommendation. Chromosome 2 displayed two linkage peaks, one at 2p21 and one at 2q14.3. The peak at 2p21, with a lod 1.831 (p= 1.8x10-3) for marker rs1377687 at 44Mb, replicated the DYX3 linkage peak, as expected since DYX3 linkage has already been confirmed in our set of Canadian families by targeted linkage analysis (Petryshen et al. 2002). History of the DYX3 locus – 2p16 vs. 2p12-p11 This study detected dyslexia linkage to 2p21, with lod scores above 1.0 from 41.871-46.278 Mb. Five independent family samples have shown positive linkage to chromosome 2p12-p16. The DYX3 locus was first identified by a genome-wide linkage study of one large Norwegian  44  family displaying autosomal dominant inheritance of dyslexia. Parametric linkage analysis with three diagnostic models returned significant lod scores at 2p15-p16 (maximum lod=4.3), and a non-parametric multipoint analysis confirmed the results (p=0.0009) and localized the linkage signal to a 4cM region, between the STR markers D2S2352 (2p16.2) and D2S1337 (2p16.1) (Fagerheim et al. 1999). This region was further refined by identification of a haplotype segregating with dyslexia in the same family (Fagerheim 2002). The first independent replication of this dyslexia locus was found in our set of Canadian families using the qualitative PCD phenotype: parametric linkage analysis showed a weak linkage signal, while NPL analysis provided significant evidence of a replicated linkage at 2p16.3 (NPL Zall score = 2.33, p=0.0087 at D2S1352) (Petryshen et al. 2000, 2002). Significant linkage signals were also obtained using multipoint variance components analysis of the quantitative measures: spelling (peak lod = 3.82, at 2p16.2-p16.1 [53.7Mb-57.1Mb]), phonological coding (peak lod = 1.13, at 2p16.1 [57.1Mb]), and phonological awareness (peak lod = 1.01, at 2p16.2 [53.7Mb]) (Petryshen et al. 2002). Fisher et al (2002a) also replicated the DYX3 signal over a 20cM interval in a set of families from the UK and USA.  These authors performed a genome-wide linkage scan using quantitative NPL for three phenotypes: single word reading, phonological awareness (p=0.0003 at 2p13 in US sample) and orthographic coding (p=0.0007 at 2p16 in UK sample). The 2p16 linkage signal was further refined to a 12 cM region between D2S337 (61Mb) and D2S286 (75Mb) in a quantitative linkage analysis with 21 microsatellite markers of the set of UK families (Francks et al 2002). A fourth group (Kaminen et al 2003) discovered linkage (NPL score = 2.55, p=0.004) to a nearby region at 2p11.2 (D2S2216 at 111cM, 88.2Mb), approximately 34cM centromeric from the linkage peak of Fagerheim et al. using a qualitative phenotype derived from performance results on Finnish reading and spelling tests. This linkage to 2p11 was replicated by Peyrard- Janvid et al. 2004. The 2p11.2 linkage signal identified by Kaminen et al. 2003, using the non- word reading phenotype, was confirmed in Dutch sibpairs by de Kovel et al. 2008 with linkage at 2p12 identified using the quantitative traits: non-word reading (lod=1.96, p=0.001 at D2S378 ) and rapid naming (lod = 1.5, p = 0.004). However, parametric analyses with various genetic models failed to produce any lod scores above 1.0 (de Kovel et al. 2008).  45  Further support for a dyslexia susceptibility gene located near 2p21 comes from the Loo et al (2004) linkage study of normal reading performance in ADHD affected sibpairs (described in greater detail in section 3.4.1).  Loo et al. reported significant evidence for linkage (multipoint maximum lod score 2.25) of a reading ability QTL at marker D2S367 (34.29 Mb) at 2p22.3 (Loo et al, 2004). All together, this research implicates a broad linkage region spanning 2p22-2p11. As mentioned previously, dyslexia mapping is complicated by the presence of diverse phenotypic measures and diagnosis criteria. The wide range of localizations for the 2p linkage signal may result from errors in parametric model specification or map positions for STR markers, which could obscure the pleiotropic effect of a single gene on different phenotypes. Alternatively, the multiple linkage peaks may reflect the presence of more than one susceptibility gene underlying the linkage signals, each found at different frequencies in the various populations studied. Support for previously identified linkage at 2q22, 7q32, 4p16-15, 17p13.3 and DYX4 My analysis revealed a slightly lower linkage peak at 2q14.3, with a lod 1.366   (p=6.0x10-3) for markers rs779976 and rs812952 (124.736 Mb and 124.738 Mb), providing some support for the 2q22 dyslexia locus that was first identified in a QTL linkage analysis in American families using the specific trait of speed of phonological decoding (Raskind et al 2005). Subsequently, Bates et al (2007) provided the first independent replication of this 2q linkage using QTL analysis for the traits regular-word spelling (peak lod 2.18 at 2q35, 217Mb) and reading accuracy (lod 1.13 at 2q14.3, 126Mb). The localization of linkage peaks for QTLs can vary significantly due to the effects of stochastic factors (Roberts et al 1999); therefore, it is likely that the peaks reported by Raskind et al (2005) and Bates et al (2007) reflect the same underlying gene identified in our NPL analysis. My NPL study also yielded a lod of 1.113 for locus 7q32.2 (at marker rs4731689, 130 Mb), which while not large enough to claim replication, provided independent support for the dyslexia locus reported at 7q32. This locus was first identified through qualitative linkage analysis in 11 Finnish families with developmental dyslexia (Kaminen et al 2003) and later replicated by QTL methods in a sample of Australian twins with normal reading and spelling ability (Bates et al  46  2007). In the same genome-wide study, Bates et al (2007) identified two novel linkages at 4p16.1-p15.33 and 17p13.3 with maximum QTL lod scores of 2.08 for irregular reading trait at 4p15.33, 1.43 for irregular spelling at 4p16.1 and 1.99 for irregular spelling at 17p13.3. The current NPL study provides some support for linkage of the PCD trait with these two loci, with lod scores of 1.028 for 4p15.1 (rs1523166, 32Mb) and 1.124 for 17p13.3 (rs8065080, 3Mb). As expected, my NPL analysis detected the DYX4 linkage signal (6q11.2-q12) that was previously identified by two-point parametric analysis, multipoint parametric analyses and non- parametric analyses in our families (Petryshen et al. 2001). The DYX4 locus has been independently replicated by a genome-wide linkage analysis of normal reading ability in Australian twins (Bates et al 2007). The current NPL score of 0.934, however, was significantly lower than the original peak NPL score of 2.21 (p=0.012) that was obtained using microsatellite markers. The reduced linkage signal probably results from a decrease in the informativeness of the current family set, since it was necessary to split some of the larger pedigrees to run the MERLIN software program (see section 2.3.2 for details). Alternatively, the linkage signal may be reduced by the use of less informative bi-allelic SNP markers instead of the polymorphic microsatellite markers used by Petryshen et al (2001) in the previous linkage analyses (see section 3.4.3 below for further discussion). 3.4.3 Failure to replicate linkage signals The current NPL analysis, using SNP markers and a qualitative PCD phenotype in our Canadian families, failed to replicate the linkage signals previously reported for the dyslexia loci DYX1 (15q21), DYX2 (6p22.2), DYX5 (3p12-q13), DYX7 (11p15.5), and DYX8 (1p36-p34) and DYX9 (Xq26-27).  Previous simulations, estimating the power of this set of Canadian families to detect linkage using parametric analysis to a PCD locus, demonstrated >86% power to generate a lod>3.0 for a major gene linked in up to 40% of families, and very good power (>72%) to generate a lod>1.0 for a gene linked in only 20% of families (Ploughman and Boehnke, 1989). It should be noted, however, that NPL analysis uses only information on allelic sharing among affected family members, while the parametric analysis also utilizes information from unaffected siblings. As already mentioned, parametric analysis has more power than NPL analysis to detect linkage if the model specified during parametric analysis is reasonable correct  47  (Abreu et al 1999). Thus, the use of NPL analysis in the current study may have reduced our power to detect the aforementioned loci. DYX1, located at 15q21, is a well-replicated dyslexia linked locus (see Table 1). Linkage between Chromosome 15 and reading and spelling disorder was first identified in a study targeted to chromosome 15 due to a centromeric heteromorphism (a natural variation in choromosome staining pattern); however, this linkage was only present in 20% of the studied families (Smith et al 1983). The DYX1 linkage (at 15q21) was not observed in a Danish study (Bisgaard et al. 1987) or in a genome-wide Finnish study (Kaminen et al 2003). DYX2 on chromosome 6p is considered the most well-replicated dyslexia locus (see Table 1 and section 4.4). However analyses with our set of Canadian families using both quantitative and qualitative analysis have consistently failed to replicate this linkage (Field and Kaplan 1998, Petryshen et al 2000). The human histocompatibility antigen (HLA) region on 6p21.3 was suggested as a potential quantitative trait locus (QTL) influencing dyslexia due to phenotypic co- morbidity between learning disabilities and immune disturbances (Smith et al. 1991, Cardon et al. 1994 and 1995). However, a shared aetiology between autoimmune disorders and developmental dyslexia has yet to be proven. Other groups have also failed to replicate DYX2 linkage (Bisgaard et al. 1987, Schulte-Körne et al. 1998, Nopola-Hemmi et al 2001, Kaminen et al 2003; Chapman et al 2004). DYX5 (3p12-q13) was identified by a genome-wide scan in a large Finnish pedigree displaying an autosomal dominant inheritance pattern (Nopola-Hemmi et al. 2001).  This linkage was later replicated by QTL linkage studies (Fisher et al 2002; Bates et al 2007). However, the current NPL analysis failed to detect any linkage signals in that region, as did Kaminen et al 2003. Suggestive linkage to Xq26-27 was first reported by Fisher et al 2002 through a multipoint QTL analysis of X-linked markers. However, the DYX9 locus was actually identified through a genome-wide scan in an extended Dutch family (de Kovel et al 2004) and then independently replicated by Bates et al 2007. My genome-wide NPL scan, as well as two other studies (Nopola-  48  Hemmi et al 2001; Kaminen et al 2003), all failed to detect linkage to the q arm of the X chromosome. Surprisingly, no linkage was detected on chromosome 11, despite the discovery of the novel DYX7 locus (11p15.5) in our same family set (Hsuing et al 2004). Hsuing et al (2004) reported significant linkage to the candidate gene DRD4, which also shows strong association with ADHD (LaHoste et al 1996; Smalley et al 1998; Swanson et al 1998; Rowe et al 1998; Sunohara et al 2000; Muglia et al 2000; Holmes et al 2000; Faraone et al 1999, 2001, 2005). Significant linkage for the qualitative PCD phenotype was found with both non-parametric and parametric analyses; however, the maximum 3-point lod score of 3.57 was attained using an autosomal dominant model with reduced penetrance (80%) (Hsuing et al 2004). The current observation of no linkage may result from the difference between non-parametric and parametric methods in power to detect linkage, especially if the applied model is in fact representative of the true mode of inheritance. While this localization of DYX7 to 11p15.5 has yet to be officially replicated, independent confirmation was provided by Fisher et al (2002a), who detected significant linkage at the most distal 11p marker they typed (D11S1338, 5Mb proximal to DRD4) . The current NPL analysis also failed to detect linkage to DYX8 on 1p36-p34, despite previous confirmation of this susceptibility locus in our 100 Canadian Families (Tzenova et al 2004). Linkage to 1p36-p34 was first reported in a targeted linkage study using the rhesus blood group (RH) locus and two other DNA markers in nine dyslexia families (Rabin et al. 1993) and, concurrently, in a German family where dyslexia and delayed speech development were co- segregating with a balanced translocation t(1;2)(p22;q31) between chromosomes 1 and 2 (Froster et al. 1993). Linkage to DYX1 has been independently replicated in four other studies, including Tzenova et al (2004), whose QTL, MLS and NPL analyses showed significant linkage to markers at 1p36, over 60 Mb distal to the current study’s observed chromosome 1 peak at 1p31.1 (lod=1.244) . The failure of the current NPL study to replicate the linkages to DYX8 and DYX7 that were previously indentified in our set of Canadian families is unexpected. A biallelic SNP marker is less informative than the polymorphic microsatellite markers used in the previous analyses. However, the SNP marker set is far denser, and multipoint analysis methods capitalize on this  49  SNP density, so the current marker set could potentially be more informative than the previous microsatellite set and enable a more precise localization of linkage signals. Comparative studies have shown that linkage analysis results using microsatellite and SNP markers are similar (Daw et al 2005; Ulgen and Li 2005). Furthermore, Lindholm et al (2004) compared the informativeness of multiple closely spaced SNPs and single microsatellites in nuclear families using an information ratio (IR; the ratio of the actual average maximum lod score to the maximum lod score attainable with a fully informative marker) and determined that a microsatellite marker with 9 equally frequent alleles returned a similar IR (86-88%) as a string of 4-5 SNPs for multipoint linkage analyses. Therefore, it is more likely that the failure to detect linkage to DYX8 and DYX7 resulted from differences in the individuals and family structures that were included in the current versus the two previous linkage analyses. As described in section 2.3.1, only a subset of the total individuals was selected to be array-genotyped, and it was also necessary to break down nine extended families into smaller pedigrees in order to run the linkage analysis using the MERLIN program (see section 2.3.2). Both of these modifications could have reduced the information on co-segregation of the genotyped markers and the PCD trait within a pedigree, thereby dampening the linkage signals. The use of NPL analysis, instead of parametric analysis, in the current study may have further reduced our power to detect the DYX8 and DYX7 loci. 3.5 Subsequent Analyses The chromosome 16p12, 2p21, and 4q13 linkage regions (which displayed the top three maximal NPL lod scores), as well as the 6p22 linkage region, were selected for further targeted genetic analysis using family based association methods. The discussion of these regions is continued in sections 4.1, 4.2, 4.3 and 4.4, respectively.      50                Figure 5: Results for non-parametric linkage analysis for chromosomes 1-22 and the X chromosome. The x-axis shows the position of markers and genetic distance in cM. The y-axis shows NPL lod scores for each position. Markers at the top of the linkage peaks with lod scores >1.0 are highlighted in color. The MERLIN NPL scoring function was used to estimate sharing of alleles over the expected value under Ho= no linkage, at each marker locus.    51                  Figure 5: continued   52                  Figure 5:  continued   53                  Figure 5: continued   54                  Figure 5: continued   55                  Figure 5: continued   56                  Figure 5:  continued   57                  Figure 5: continued   58                  Figure 5: continued   59                  Figure 5: continued   60                  Figure 5: continued   61           Figure 5: continued  62  CHAPTER 4 FAMILY-BASED ASSOCIATION ANALYSIS The three regions 16p13-p12, 2p21 and 4q12-q13 were selected for further genetic analysis using the method of family-based association testing to search for evidence of linkage disequilibrium between a dyslexia locus and regional markers. These regions displayed the three highest maximal lod scores from the NPL analysis, as summarized in Chapter three. The limits for the three search regions, which centered under the three linkage peaks, were established on either side of the maximum lod score at the point where the lod scores fell below a cutoff value of lod=1.0. A fourth region, at 6p22.2, has been reported linked to and associated with dyslexia; we therefore targeted it for association analysis to see if association could be detected, despite the lack of any linkage signal in our families. It is known that association analysis is a more sensitive method for detecting susceptibility for complex trait genes than linkage analysis (Risch and Merikangas, 1996). These regions were tested for association with a dense marker map in a set of 83 Canadian trios. Because of this modest sample size, the detection of association may be limited to alleles that have a stronger genetic effect and are at a higher frequency in the study sample. Many of the markers were correlated due to their LD (r2≤0.8), so the multiple tests for association between each marker and the PCD trait are not independent. Furthermore, there is a higher likelihood of finding association in a region with ‘a priori’ evidence for linkage, since the presence of a dyslexia susceptibility gene is already indicated. The Bonferroni correction for multiple testing, based on the assumption of independent tests, is therefore far too conservative for these analyses. In the following analyses, the null hypothesis of no association was rejected if the pointwise p-value was smaller than or equal to a significance level of 0.05. Searching for Candidate Genes Developmental dyslexia research frequently reports abnormalities in brain development of dyslexic subjects that, according to the phonological theory, translate into problems in phonological processing. Neurons are initially produced in the centre of the developing brain and rely on proper chemical communication to migrate out, usually along glial fibers, towards the cortex or other destinations in the brain. Aberrant cell migration in the brain can result from mutant molecules involved in the guidance of neuronal migration or the stability and remodeling  63  of the cytoskeleton (Pilz et al 2002).  Once migrating neurons have reached their destination, they must also extend their axons and dendrites. Alterations in this neuronal pathfinding are another possible neurodevelopmental mechanism for developmental dyslexia that may account for the abnormal axonal connections between the right and left hemispheres of the brain in dyslexic individuals (Galaburda 1989; Galaburda et al 2006). In the search for dyslexia genes with a role in neuronal migration or axonal guidance, we can restrict our selection to genes for which the protein products are expressed ubiquitously or predominantly in the brain. Genes that encode receptors, adhesion molecules, extracellular matrix proteins, transcription factors, intracellular signaling molecules and members of signaling cascades are all likely candidates (Kriegstein and Noctor 2004). For example, the DYX5 candidate gene ROBO1 is an axon guidance receptor gene that encodes a receptor found on the surface of the axon growth cone (Hannula-Jouppi et al 2005). Mutations in robo (the Drosophila homologue) resulted in the improper crossing of axons between the right and left hemispheres of the brain (Seeger et al 1993; Kidd et al 1998). Gene knockout mouse models for Robo1 have confirmed that this receptor plays a role in the regulation of tangential cell migration in the brain (Hu 1999; Wu et al 1999, Zhu et al 1999; Andrews 2006). The dyslexia-associated DYX1C1 gene at 15q21 encodes a protein that is likely involved in neuronal migration (Taipale et al 2003); further evidence comes from an experiment where the inhibition of Dyx1c1 (the rat homologue) expression in rat embryos disrupted neuronal migration in the developing neocortex and resulted in anatomical changes, as well as altered learning and motor skills (Frenkel et al 2000). 4.1 Testing for Association at 16p12 4.1.1 Introduction A novel dyslexia-linked locus at 16p12 displayed the strongest linkage signal (NPL lod 2.263, p=6.2x10-4) in my genome-wide scan and was, therefore, of high priority for subsequent genetic investigations by association methods to detect dyslexia susceptibility genes. This region has also been reported to show linkage to ADHD in the presence of reading disability (Loo et al 2004). While no studies have examined 16p12 for an association with dyslexia, markers in this  64  region have recently been associated with ADHD and autism in positional candidate gene studies (Turic et al 2004; Barnby et al 2005).  An analysis of 239 multiplex families from the International Molecular Genetics Study of Autism Consortium examined the coding variants in 10 positional candidate genes at 16p13-p11 and found evidence (Fisher’s exact test, p<0.001) of single-marker association of autism with variants in three of the ten genes examined: ABAT, CREBBP and GRIN2A (Barnby et al 2005). Turic et al (2004) also detected significant association within exon 2 of the GRIN2A gene (χ2 = 5.7, p=0.01) in a family-based association analysis of 238 families displaying heritable ADHD. 16p12 has also been reported to harbour a risk gene for bipolar disorder in a study conducted by the Wellcome Trust Case Control Consortium (2007) that identified a strong association for marker rs420259 at 23.54Mb (genotypic test p = 6.29x10-8) in a case-control study with 2000 cases and 3000 controls. Among the newer genetic discoveries, an unexpected revelation is that the genomes of normal individuals contain a large number of copy number polymorphisms (CNPs, also known as copy number variants or CNVs) that are spread throughout euchromatic and pericentromeric DNA (Iafrate et al 2004; Sebat et al 2004). These recently-evolved large structural polymorphisms are not only examples of human genome variation, but also serve as a possible mechanism for the differences observed in human disease susceptibility. The Database of Genomic Variants documents a number of structural and copy number variants (CNVs) that are found within 16p12-p11 and result in gene content differences among humans in the general population. Chromosome 16 displays one of the highest levels of segmental duplications of all the human autosomes, as revealed when Martin et al (2004) finished sequencing 78,884,754 base pairs (over 99.9%) of its euchromatin. CNVs (genomic microduplications and microdeletions) have recently been identified as important contributors to disease risk, especially for neurological and psychiatric disorders (Lee and Lupski 2006; Sebat et al 2007; Kumar et al 2008; Walsh et al 2008; Weiss et al 2008). Of particular importance is the reported association between recurrent 500kb microdeletions and microduplications at 16p11.2 with schizophrenia (Walsh et al 2008) and autism (Weiss et al 2008; Kumar et al 2008). The breakpoints of this 500kb deletion have been mapped to the edges of flanking 147kb segmental duplications (Kumar et al 2008); the more distal breakpoint at ~29.5Mb is located just 6Mb proximal to the region targeted in the current association analysis for PCD.  65  4.1.2 Method An 8.3 Mb region at 16p13.11-p12.1, from 15.684 – 23.978 Mb, was investigated for dyslexia candidate genes by testing for association between 176 tag SNP markers and the PCD phenotype in 83 simplex trios (see section 2.3.3 and 2.8.1 for description of family set and marker selection). Genotyped marker data were analyzed for single-marker and haplotype (two marker sliding window) associations using the TDT in PLINK, while the AFBAC program was also used as an independent method to test for single-marker association (see section 2.8.2 for more information on these statistical methods of analysis). Markers and candidate genes were annotated using the UCSC genome browser (http://genome.ucsc.edu/) on the Human Mar. 2006 Assembly (Kent et al 2002; Karolchik et al 2008). 4.1.3 Results Eleven markers displayed association with a pointwise significance level of p<0.05, for either the AFBAC or TDT single marker association tests (see Table 6). The 10 markers that obtained a nominal p<0.05 for both the AFBAC test and TDT were considered to be significantly associated with the PCD trait (highlighted in blue in Table 6). All SNPs that displayed single marker association also displayed a significant 2-marker haplotype association, when that SNP was analyzed with a flanking SNP found on either side by sliding window analysis.  Table 6 lists the lowest p value obtained for a marker in a specific 2- SNP haplotype when the TDT was used to test for haplotype association with PCD. The p values for the rest of the haplotype combinations in a window, for all markers that displayed significant single-marker association, are listed in Table 13 in Appendix 1. 4.1.4 Discussion Results from the TDT and AFBAC test identified 10 markers of the 176 tested at 16p12.3- 12.1 that were significantly associated with PCD. The search for possible candidate genes (listed in Table 7A) encompassed a region of approximately 6.7Mb covered by these markers, from the most distal marker (rs9934960 at 17.161049Mb) to the most proximal marker (rs120908 at  66  23.871593Mb). 16p12.3-p12.1 is a particularly gene-dense region of the genome, with a number of genes that are highly expressed in the brain or expressed ubiquitously. As well, there are four CNPs, listed in Table 7B, that further complicate the genomic structure of the 16p12.3-p12.1 region. Over sixty Refseq genes and UCSC predicted genes reside at 16p12.3-p12.1 (see Table 7A); genes with positional support, based on the associated marker data, were selected as possible candidates, before considering the biological support. A marker identified as significantly associated at p<0.05 may well represent a false positive result. If a causal variant has been identified (or as is more likely, if a marker locus is close enough to the disease locus for a marker allele to be associated with the disease allele), then one would also expect to see a positive association for other neighboring markers that are in appreciable linkage disequilibrium with this causal variant/associated marker. Of the 10 significantly associated SNPs, in two instances, two markers were located in the same gene. Markers rs12447418 and rs12927277 are 4.123kb apart and located in intron 3 and intron 4, respectively, of the CP110 gene. The second set of markers, rs9940479 and rs120908, are only 88bp apart and lie in intron 2 of the gene PRCKB1.  A third instance of two markers in the same location showing association to PCD are markers rs9939346 and rs1052112, separated by 13.868kb and lying in non-coding DNA between the genes KIAA0220 and H23ST2. A pair of associated markers is more likely than a single associated marker to represent a true association signal. These three sets of pairs are discussed immediately below. CP110 as a Dyslexia Candidate Gene CP110 at 16p12.3 encodes a coiled-coil centrosomal protein, necessary for centriole duplication, which is ubiquitously expressed, with elevated expression in the testis (Chen 2002). The centrosome is the microtubule-organizing centre in animal cells and is involved in cell-cycle control, organization of the mitotic spindle poles, cell motility and adhesion. The centrosome is also implicated in the first two steps of the 3-step process of cell migration: leading-edge extension, nuclear translocation and retraction of the lagging edge (Gotlieb et al 1981; Morris et al 1998; Ueda et al 1997; Hatten 2002). Centrioles give rise to the centrosome, but when they are tethered to a membrane in the cell cortex they convert into basal bodies and can form cilia  67  (Pearson et al 2007). The CP110 protein, along with CEP97, forms a centriolar complex that inhibits the conversion of centrioles to basal bodies, thereby preventing cilia formation through a form of negative regulation (Spektor et al 2007). Cilia are known for their role in cell motility and sensory perception; however, they are also involved in developmental processes such as the propagation of morphogenetic signals in embryogenesis (Badano et al 2005). Loss-of-function mutations that lead to basal body dysfunction are responsible for a variety of disorders including the reversal or randomization in body symmetry (Yokoyama et al 1993; Morgan et al 2002; Ansley et al 2003). Proteomic analysis (Anderson et al 2003) has identified variants in centrosomal proteins responsible for a range of genetic disorders including schizophrenia, where the DISC1 gene is involved in neurite outgrowth (Miyoshi et al 2004), and lissencephaly (smooth brain), which is caused by mutations in the doublecortin (DCX) gene and lissencephaly 1 gene (LIS1). DCX and LIS1 protein products localize to the centrosomal regions and are involved in neuronal migration (Morris et al 1998; Tanaka et al 2004). While at first glance, a potential link between the CP100 gene and dyslexia might be overlooked, it is actually a very strong biological dyslexia candidate gene with a possible role in neuronal migration. PRKCB1 as a Dyslexia Candidate Gene The SNPs rs9940479 and rs120908 lie in intron 2 of the gene PRKCB1 at 16p12.1, which, through alternate splicing, codes for both isoforms of the protein kinase C β enzyme. Protein kinase C (PKC) is a family of phospholipid-dependent enzymes that are involved in the regulation of gene expression, signal transduction, and control of cell differentiation and division.  The PKC family acts as a central regulator of many stress response pathways in the cell by initiating downstream phosphorylation cascades that alter trafficking pathways (Burston H; personal communication). The activation of PKC enzymes, which is often followed by translocation of the enzyme from the cytosol to the plasma membrane, can trigger increased neurotransmitter release from hippocampal neurons (Majewski and Iannazzo 1998) and influence neuronal plasticity and early-stage hippocampal long term potentiation (Ben-Ari et al 1992). The two PRKCB1 isoforms are expressed predominantly in brain structures (striatum, suprachiasmatic nucleus, and granule cells in the cerebellum and in the dendate gyrus of the hippocampus), but are also expressed in the digestive tract, immune system and kidney (Lintas et  68  al 2008). Within the granule cells (tiny neurons) of the hippocampus, the concentration of PKCβ1 isoenzyme has been shown to influence memory, circadian rhythms, and learning in both rats and mice (Columbo et al 1997; Bult et al 2001; Wu et al 2007). Wu et al 2007 demonstrated that passive avoidance conditioning in normal rats induced PKCβ1 translocation from the cytosol to the membrane, while prenatally-stressed rats, demonstrating classical learning and memory impairments, showed decreased expression of PKCβ1 protein and mRNA and no PKCβ1 translocation when these rats were exposed to the same passive avoidance training. Evidence in human studies also demonstrates a neurobiological role for PRKCB1 (Philippi et al 2005; Lintas et al 2008).  Several single SNPs (most significant p=0.0027), as well as a 3SNP- haplotype (p=2.96x10-5), within the PRKCB1 gene were associated with autism in 116 families, followed by replication of these associations in a second set of 167 trios (Philippi et al 2005). The 3 SNPs in the autism-associated haplotype (rs3785387, rs196002, and rs1873423) were all located in intron 2 of PRKCB1, which is the same intron that holds the two SNPs (rs9940479 and rs120908) that show significant association with PCD in my analysis.  Another family-based association study, designed as an independent replication of the Phillipi et al 2005 study, reported both single-marker (p=0.04) and a 4-SNP haplotype (p=0.0078) association with autism in 229 simplex and 5 multiplex families (Lintas et al 2008). The 4 SNPs were also located in intron 2 of PRKCB1, and this autism-associated haplotype was correlated with PRKCB1 gene expression in postmortem autistic brain samples; a significant reduction in PRKCB1 mRNA and protein levels was observed in samples from the superior temporal gyrus of 11 patients compared with matched control subjects (Lintas et al 2008). The genetic analyses and neurophysiological evidence suggest that PRKCB1 is a strong candidate gene for dyslexia, a gene that may encode an intracellular signaling molecule with a role in the regulation of synaptic transmission. PRKCB1 may initiate a chain of phosphorylation events that trigger the release of neurotransmitters, which could affect neuronal migration and axonal pathfinding during brain development or mediate the neurobiochemical modifications of learning.   69  A Possible Role in Dyslexia for Low-Copy Repeats that Resulted From Human-Specific Expansion Markers rs9939346 and rs1052112 are located in the non-coding DNA between the genes KIAA0220 and H23ST2. The protein encoded by Heparan sulfate D-glucosaminyl (HS3ST2) belongs to the heparan sulfate biosynthetic enzyme family, also known as 3-O-sulfotransferases (3OSTs), and is a type II integral membrane protein (Borjigin et al 2003). This gene is expressed predominantly in the brain and may play a role in the nervous system: it is highly expressed in the rat pineal during daylight hours and is involved in circadian rhythm (Borjigin et al 2003; Lawrence et al 2007).  KIAA0220 encodes a hypothetical protein with a nuclear pore complex interacting protein (NPIP) domain and may represent a gene or gene fragment. It belongs to the putative KIAA0220/SMG1 gene family and is within a block of low-copy repeats found on chromosome 16, called LCR16u (Johnson et al 2006). As mentioned in section 4.1.1, there is a 500kb interval at 16p11.2 with recurrent CNVs that have been associated with schizophrenia and autism. This interval holds approximately 54 intrachromosomal duplications, making it one of the most highly duplicated regions on chromosome 16 (Martin et al 2004). In humans, chromosome 16 euchromatin has 17 complex blocks of low-copy repeats (known as LCR16) that contain various gene and gene fragments, including NPIP and KIAA0220/SMG1 (Johnson et al 2006). In particular, there is a complex distribution of segmental duplications (4.2Mb of sequence) that seem to have played a role in hominoid evolution (Eichler et al 2001; Martin et al 2004; Johnson et al 2006). Comparative genomic hybridization analyses have revealed that LCR16a duplication began early during primate evolution (before the human/macaque lineages diverged) and continued to expand on chromosome 16 independently in the human lineage (after the separation of humans and chimpanzees) (Johnson et al 2001; Johnson et al 2006). LCR16u, a second LCR, was duplicated in concert with LCR16a, even though they descended from different ancestral loci (Eichler et al 2001; Johnson et al 2001). LCR16u contains the KIAA0220/SMG1 putative gene family and is found approximately 8 times in the human genome. LCR16a, which has 17 copies, contains one of the most rapidly evolving hominoid gene families, called morpheus, which includes the NPIP gene (nuclear pore  70  complex interacting protein; accession AF132984). Exons 2 and 4 from the genes of this family have shown especially significant positive selection (Eichler et al 2001). The current genetic analysis of dyslexia was not designed to investigate the complex genetic architecture at 16p12.1-p11.2, although it is intriguing to note the detection of an association between dyslexia (a disorder unique to humans) and SNPs in a region rife with segmental duplications that are specific to hominoid evolution. It would be unwise to attempt to draw conclusions from a study with a limited sample in an area which was not the focus of the research, even if the results suggest that the area merits further investigation. In contrast, however, the current association scan has been successful in identifying CP110 and PRKCB1 as two strong candidates for dyslexia susceptibility genes. 4.2 Testing for Association at 2p21 (DYX3) 4.2.1 Introduction DYX3 is a widely-replicated dyslexia locus that has shown positive linkage and association signals for dyslexia in many different populations.  All together, however, this research implicates a very large region on chromosome 2 from 2p21-p12; more than one susceptibility locus, therefore, may be responsible for the various signals detected across the p arm of chromosome 2. My NPL scan identified a peak (lod = 1.831) at 2p21 (44.195Mb) and targeted a 4.4 Mb region for further analysis. A fine-mapping association study was conducted, using a dense set of SNP markers, to further refine this region and to identify a candidate dyslexia susceptibility gene for DYX3. A Scandinavian research group, who discovered linkage to 2p11 in Finnish pedigrees (Kaminen et al 2003; Peyrard-Janvid et al 2004), has performed association analyses in an attempt to identify candidate genes underlying the dyslexia 2p11 linkage peak. A haplotype pattern mining method (Toivonen et al 2000) was used to perform case-control association on a set of independent family trios and identified significant association with markers D2S286/rs3220265 (p<0.001) (Peyrard-Janvid et al. 2004). When this 670kb region was further characterized through the addition of 6 SNPs, a haplotype association pattern was no longer observed and the TACR1 (tachykinin receptor 1), located at marker D2S286, was excluded as a  71  candidate gene (Peyrard-Janvid et al. 2004). Further linkage disequilibrium mapping using the TDT in the set of Finnish families, as well as independent replication in 251 German families, identified two overlapping risk haplotypes spanning 16kb in an intergenic region between the genes FLJ13391 and MRPL19/C2ORF3 (Anthoni et al 2007). While the coding variants in these two genes did not show significant association with dyslexia, the expression pattern for these two genes correlated with the other four dyslexia candidate genes (DYX1C1, ROBO1, DCDC2, and KIAA0319) (Anthoni et al 2007). Association mapping of the 2p region has also been performed in the USA population. After refining the 2p linkage signal with a quantitative linkage analysis for a set of Colorado dyslexia sib-pair families, Francks et al (2002) continued with a sibpair quantitative association analysis, using the same set of 21 microsatellite markers in the 2p12-p21 region, and found suggestive association for two markers with p values < 0.05. The 2p12 region (marker D2S2114 at 75.684Mb) and the 2p21 region (marker D2S2378 at 46.112Mb, within the PRKCE gene) returned empirical p values of 0.004 for both word-recognition and phonological decoding measures (Francks et al 2002). Two positional candidate genes were screened (SEMA4F and OTX1), but association testing revealed that the 2p12 linkage signal was not due to coding variants in these genes. Of special interest is finding of association with the phonological decoding phenotype for marker D2S2378 (Franks et al 2002), since this marker falls right in the 2p21 region that was targeted for association analysis through my NPL linkage scan (see section 3.4.2), with a peak lod score of 1.831 for SNP marker rs1377687, located at 44.195480Mb. 4.2.2 Methods and Results A 4.4 Mb region at 2p21, from 41.871 – 46.278 Mb, was tested for association between 185 tag SNP markers and the PCD phenotype in 83 simplex trios (see section 2.3.3 and 2.8.1 for description of family set and marker selection), according to the same method applied in section 4.1.2. Fourteen markers displayed association, with a pointwise significance level of p<0.05, for either the AFBAC or TDT single marker association tests; however, only 10 markers that showed significant association with the TDT also maintained a pointwise p<0.05 for the AFBAC test (these 10 markers are highlighted in blue in Table 8). All SNPs that were associated with the  72  PCD trait in the single marker tests also displayed a significant 2-SNP haplotype association. Table 8 lists the lowest haplotype TDT p-value returned for a marker, out of all haplotype combinations tested for that marker. The haplotype TDT results for all other haplotypes that include these markers are listed in Table 14 in Appendix 2. 4.2.3 Discussion Ten markers, out of the 185 tested, showed significant association with the qualitative PCD phenotype by both AFBAC and the TDT methods. The region tested for association spanned 4.4Mb, from 41.871-46.278Mb, but the 10 associated markers fell into a 2.8Mb region, from 43.105656Mb-45.947264Mb, that was examined for candidate genes using the UCSC genome browser. These genes are listed in Table 9. Three markers in particular (rs4557033, rs4953273, and rs10189339) were all located in the same gene, PRKCE (protein kinase C epsilon). The first marker resides in intron 1, while the other two markers reside in intron 2. Two other PRKCE markers (rs10865208 and rs13036100) were significant only by TDT single marker and haplotype tests (see Table 8). The microsatellite marker D2S2378 is also located in the PRKCE gene (intron 10); D2S2378 was associated with the phonological decoding phenotype by Francks and colleagues (2002), as mentioned in section 4.2.1. Markers rs17335631 and rs6748247 were located in the gene PLEKHH2; however, this gene was not considered a good candidate since its expression pattern is predominantly non-brain (according to microarray expression data from GNF Atlas 2 on the UCSC).  PRKCE (protein kinase C epsilon) encodes the PKCε enzyme, which is a member of the same family of phosphorylating PKCs described in section 4.1.4 in relation to dyslexia association with PRKCB1 markers on chromosome 16p12.  Members of the PKC family are involved in a range of cellular signaling pathways, with different outcomes depending on the type of cell. Accordingly, members of the PKC family show tissue-specific distribution and different expression profiles: PRKCB1 is predominantly expressed in the brain, while PRKCE is expressed ubiquitously (Decker and Parker 1994). PRKCB1 is a classical Ca2+ -dependent PKC, while PRKCE is Ca2+ -insensitive; however, both isoforms are activated by diacylglycerol and phorbol esters (Fagerström et al 1996). The fact that positive dyslexia associations were seen for  73  markers in two members of the same gene family (at 2p21 for PRKCE and at 16p12.1 for PRKCB1, see section 4.1.4) provides further support that the detected association signals are biologically relevant. PRKCE has been implicated in a number of different cellular functions, but its involvement in neurite outgrowth, cell signalling, neuron channel activation, and controlling anxiety-like behaviour in knock-out mice, has been revealed in experiments by Fagerström et al (1996), Ivaska et al (2002), and Hodge et al (2002). These cellular functions are of particular interest for implicating this gene as a dyslexia candidate. Wild-type human SH-SY5Y neuroblastoma cells can be induced to differentiate and display a sympathetic neuronal phenotype, and they can also be induced to form neurites (elongated, membrane-enclosed protrusions of the cytoplasm that are a precursor to the formation of functional axons) (Fagerström et al 1996). Fagerström et al (1996) reported that the application of specific PKC inhibitors prevented SH-SY5Y cells from developing the differentiated phenotype and, furthermore, that the activation of PKCε is necessary for neurite outgrowth and maintenance of the growth cones structure in SH-SY5Y cells.  Ivaska et al (2002) have demonstrated a role for PKCε in cell migration. When a cell receives a signal to migrate, integrins (receptors that facilitate cell migration and extracellular matrix attachment) are sent to the cell surface. Ivaska et al (2002) concluded that the localization of B-1 integrin at the cell surface is regulated by PKCε, and that PKCε and B-1 integrin co- localize at the surface of migrating cells. They found, however, that B-1 integrin became trapped in internal cellular compartments when PKCε was knocked out or when its catalytic activity was inhibited, suggesting that PKCε is required for exit of B1 integrin from one of these compartments back to the surface. When B-1 integrin was trapped in these compartments, cell migration was impaired. Ivaska and colleagues (2002) proposed that PKCε might be required for the formation of vesicles that carry B1 integrin back to the surface from the internal compartments. Hodge et al (2002) reported that PKCε also regulates enhancers of the GABAA receptors in the brain and, therefore, PKCε can inhibit the release of neurotransmitters by hyperpolarizing the neuronal membrane. GABAA receptors act as a chloride channel and allow the negative ion to enter a cell when they are activated by the inhibitory neurotransmitter GABA (gamma-  74  aminobutyric acid). PKCε knockout mice, which are supersensitive to the activity of enhancers for GABAA receptors, showed reduced levels of anxiety-like behavior and decreased levels of stress hormones (Hodge et al 2002). The stress hormone levels returned to normal, however, when GABAA receptor antagonists were given to the PKCε knockout mice, allowing the mice to display normal anxiety-like behaviour during two standard stress tests. The above studies provide strong evidence that PKCε is involved in intracellular signaling in neurons.  The ability of PKCε to mitigate neurite formation, cell migration, and the release of behaviour-altering neurotransmitters makes the PRKCE gene an excellent candidate for developmental dyslexia. The concentration of three significantly associated markers in the PRKCE gene, therefore, is a strongly positive result of this research. 4.3 Testing for Association at 4q12-q13.1 4.3.1 Introduction  My genome-wide NPL scan identified a novel linkage at 4q12-13.1 (described in section 3.4.1).  Chromosome 4 has not been previously identified as a dyslexia-linked region and, consequently, it has never been tested for association to dyslexia. However, a number of quantitative trait loci (QTLs) have been identified on chromosome 4 for general cognitive ability in Caucasian children in a pooled case/control study (Fisher et al 1999).  Genetic mapping for the general cognitive ability (GCA) phenotype, measured as a composite score across a variety of cognitive tests, should identify the common genetic loci that influence different brain functions, such as memory and learning, involved in cognitive functioning. Eleven microsatellite markers (out of 147 tested) showed significant association with GCA in a sample of 51 children with high cognitive ability compared to 51 controls with average cognitive ability.  Three of these associations (for markers MSX1 at 4p16.2, D4S1607 at 4q34.3, and D4S2943 at 4q35.1) were replicated in an independent set of 50 cases and 50 controls (Fisher et al 1999). A variety of brain and cognitive endophenotypes (identified through volumetric brain MRI and cognitive test performance) have been mapped to multiple regions on chromosome 4 in a genome-wide linkage and family-based association analysis of 705 individuals from the Framingham Original and Offspring cohorts, a community-based sample from Massachusetts,  75  USA (Seshadri et al 2007). These heritable endophenotypes were associated with SNPs (genotyped on Affymetrix 100K Human Gene Chip Arrays) within numerous biologically relevant genes – that is, genes that have been previously-associated with brain-based disorders including Alzheimer’s disease, schizophrenia, mood disorders, and dyslexia. Four different phenotypic measures pointed to regions of the genome that are of interest for dyslexia. Both the total cerebral brain volume (measured by MRI) and a cognitive test of attention and executive function were associated with markers in the dyslexia candidate gene DCDC2 (6p22.2, DYX2) (with p values of 6.0x10-4and 9.5x10-4, respectively), while cognitive performance on the Wide- Range Achievement Test was linked (lod 5.10) to the dyslexia locus DYX6 at 18p11.2 (Seshadri et al 2007).  Of particular interest was their detected linkage (p=3.1x10-5) of parietal brain volume to marker rs1472962 on the q arm of chromosome 4, distal from the region targeted for the current association analysis at 4q12. 4.3.2 Methods and Results A 5 Mb region at 4q12-q13.1, from 57.177–62.070 Mb, was investigated for putative dyslexia gene candidates by testing for association between 224 tag SNP markers and the PCD phenotype in 83 simplex trios (see section 2.3.3 and 2.8.1 for description of family set and marker selection). Genotyped marker data was analyzed for single-marker and haplotype (two marker sliding window) associations as previously described in section 4.1.2. Sixteen markers displayed association, with a pointwise significance level of p<0.05, for either the AFBAC or TDT single marker association tests (see Table 10). Only 13 markers with a p<0.05 for both the AFBAC test and the TDT were considered to be significantly associated with the PCD trait (these markers are highlighted in blue in Table 10). Fourteen of the 16 SNPs that displayed single marker association also displayed a nominal p-value < 0.05 in the 2 marker haplotype association. Table 10 lists the lowest p-value obtained for the specified marker when it was tested for 2-marker haplotype association with PCD, irrespective of which other marker was involved. The p-values generated for all of the allelic combinations of the two markers in the window are listed in Table 15 in Appendix 3.   76  4.3.3 Discussion Thirteen markers in the 4q12-q13 region were associated with the PCD trait, displaying a nominal p<0.05 for both the AFBAC test and the TDT. All significant SNPs were annotated using the UCSC genome browser http://genome.ucsc.edu/ (Kent et al 2002; Karolchik et al 2008). These 13 markers flanked a region spanning ~4.7Mb, from 57.291670-61.950748Mb, that was examined for possible candidate genes, listed in Table 11A. Only one of the 13 associated SNPs (rs17238975), however, was located within a gene (LPHN3). Five genes are located at the distal end of this region, but are unlikely candidates, since they display low expression in both adult brain structures and the fetal brain (according to microarray expression data from GNF Atlas 1 and Atlas 2 on the UCSC). On the other hand, lactrophilin 3 precursor (LPHN3) is an excellent candidate gene, given that its translated protein product is a neuronal receptor that is highly expressed both in adult brain structures and the fetal brain. LPHN3 has both a shorter sequence (575,330bp RefSeq with 24 coding exons, NCBI Accession NM_015236.4) and a longer predicted sequence (870,794bp, with 26 exons and 24 coding exons), which begins at 61,749,970bp and encompasses the PCD-associated marker rs17238975 at 61,950,748 bp. The presence of two nearby copy number polymorphisms (CNPs) in this region (listed in Table 11B) could also alter the expression of LPHN3; cnp344 and cnp345 are possible markers for future association analyses with PCD in our families. LPHN3 encodes a transmembrane protein that belongs to the secretin/calcitonin family of G protein-coupled receptors – a family that tends to regulate secretion processes (Lelianova et al 1997; Krasnoperov et al 1997). The physiological function of the lactrophilins (LPH) is still unknown and their endogenous ligands and intracellular signal transducing proteins have yet to be characterized. However, lactrophilins are likely to act as cell adhesion molecules coupled to signal transduction, since their intracellular C terminus binds PDZ-domain proteins called Shanks that function as molecular scaffolds and target LPHs to the synapse (Kreienkamp et al 2000). LPH, also known as CIRL (Ca2+ -independent receptor for LTX), was first discovered in an attempt to identify a calcium-independent receptor for α-latrotoxin (αLTX), a component of the  77  black widow spider’s venom that is toxic to vertebrates (Davletov et al 1996; Krasnoperov et al 1996). Neurexin (NRX), a neuronal single-transmembrane receptor, was previously discovered as a calcium dependent receptor for αLTX (Ushkaryov et al 1992). Latrotoxins are tissue- specific compounds that only act on neuronal cells; their toxicity results from their ability to trigger the exocytosis of neurotransmitters upon binding to specific neuronal receptors (Ushkaryov et al 2004). There are three isoforms of LPH; however, LPH1 is the only homologue that has a high affinity for αLTX (Ushkaryov et al 2004). The three isoforms of LPH display different expression patterns: LPHN3 is brain-specific, while LPHN1 and LPHN2 are ubiquitous, but with opposite patterns - LPHN1 is expressed predominantly in the brain and LPHN2 is only found at low levels in the brain (Sugita et al 1998; Ichtchenko et al 1999; Matsushita et al 1999). The expression of these three isoforms is highly regulated during postnatal brain development in rats: LPHN3 expression peaks just after birth, followed by a peak expression of LPHN2 a few days later, while LPHN1 mRNA expression is inversely proportional with age, with highest expression occurring in the aged rat (Kreienkamp et al 2000). The highly regulated temporal expression of the LPH isoforms is indicative of a role in synaptogenesis. LPH and NRX, as neuronal cell surface receptors involved in the regulation of presynaptic release of the glutamate and GABA neurotransmitters, are compelling biological candidates for brain-based disorders. Interestingly, the neurexin1 gene (NRXN1), at 2p16, which codes for one of the multiple forms of NRXs, has been associated with autism (Feng et al 2006; Szatmari et al 2007; Kim et al 2008). Furthermore, LPH3 and LPH1 have exhibited differential expression in rat CA1 and CA2 hippocampal neurons when the supply of oxygen and glucose is reduced in vivo (Bin Sun et al 2002), suggesting that LPH3 and LPH1 may moderate the increased vulnerability of neurons in the CA1 region of the hippocampus (compared to neurons in the CA3 region) to the neurodegeneration that results from cerebral ischemia. The power of our 83 trios to detect an association was limited by the small sample size. However, my analyses did identify two PCD-associated SNPs (with p-values of 0.006 for marker rs7690173 and 0.007 for rs10517524) just upstream of a plausible candidate gene, LPHN3, as  78  well as a weaker signal (p=0.04) within the gene. Given the neurobiology of the LPHN3 protein product at the synapse and its role in neurotransmitter release, this region deserves further study, including an association study with nearby CNVs. 4.4 Testing for Association at 6p22.2 (DYX2) 4.4.1 Introduction A dyslexia susceptibility locus, called DYX2 and fine mapped to 6p22.2, has been identified and replicated in a number of independent samples. Various groups have reported linkage of dyslexia to chromosome 6p23-p21.3 near HLA (Smith et al. 1991; Cardon et al. 1994, 1995; Gayán et al. 1999; Knopik et al. 2002; Kaplan et al. 2002; Grigorenko et al. 1997, 2000, 2003; Fisher et al.1999, 2002; Francks et al 2004; Deffenbacher et al 2004). Attempts to identify the genes responsible through linkage disequilibrium mapping studies, however, have produced positive results for association between dyslexia and two gene clusters in this region: the VMP/DCDC2/KAAG1 locus (Deffenbacher et al. 2004, Meng et al. 2005, Schumacher et al. 2006) and the KIAA0319/TTRAP/THEM2 locus (Francks et al. 2004, Cope et al. 2005, Harold et al 2006, Paracchini et al 2006). The distal cluster contains a Vesicular membrane protein p24 gene (VMP), a Doublecortin-domain-containing-2 gene (DCDC2) and a Kidney-associated- antigen-1 gene (KAAG1). The proximal cluster contains KIAA03109, a TRAF-and-TNF- receptor-associated-protein gene (TTRAP) and a Thioesterase-superfamily-member-2 gene (THEM2). No evidence for linkage to this region was found in our family set in the current NPL analysis, as seen in section 3.3, or in previous studies of our families using both qualitative parametric and non-parametric analysis as well as quantitative (QTL) analysis to measures of phonological coding, phonological awareness, spelling, and RAN speed (Field and Kaplan 1998; Petryshen et al. 2000). Other dyslexia studies have also failed to detect linkage to 6p22 (Nothen et al. 1999; Chapman et al. 2004; Bates et al. 2007, Brkanac et al 2007; de Kovel et al. 2008), suggesting that DYX2 predisposition may vary between populations or only be pertinent to a sub-sample of the phenotypic measures. Power calculations have shown that NPL methods are highly effective for identifying a major locus with a large effect on genetic risk, but when there  79  are a number of predisposing genes with small genetic effects, then NPL methods require an inordinately large sample size to compensate for the reduction in excess allele sharing (Jones 1999). In these cases, methods that search for allelic associations (linkage disequilibrium) require a much smaller sample size (Jones 1999). If DYX2 acts as a susceptibility locus with a small genetic effect in our families, then the sample size in my NPL scan may have been too small to detect linkage at this locus. Consequently, a 700kb region encompassing the two gene clusters previously reported to be linked and associated with dyslexia was tested for association with PCD status in a set of 83 simplex trios to see whether LD fine-mapping of this area would return a positive result or support the conclusion of no or little DYX2 effect in our Canadian families. 4.4.2 Methods  Twenty-three markers were tested for association with qualitative PCD affection status in a set of 83 simplex trios selected from our 100 Canadian families according to selection criteria described in section 2.3.3. The 23 tag SNP markers (identified by measures described in section 2.8.1) encompassed a 700kb region on chromosome 6p, from 24.200-24.900 Mb, with an average intermarker distance of 30.3kb. These markers came from a larger set of markers which were in HWE and were genotyped in over 98% of individuals (as described in section 2.6.3). The TDT and AFBAC tests were used to search for any single marker association with PCD (section 2.8.2), while the TDT was used to test for any two locus haplotype association with PCD with a sliding window analysis (section 2.8.3). 4.4.3 Results All of the 23 markers tested failed to show association with PCD (p>0.05), as shown in Table 12, for both the AFBAC and TDT single marker association tests. The 2-locus haplotype analysis generated 22 sliding windows producing 73 two-marker allelic combinations that were tested for association by the TDT. Only one allelic combination between the major allele for marker rs16888748 and the minor allele for marker rs3804320 displayed significant under-transmission compared with the 50% expected transmission (p=0.0196). However, neither of these two markers showed significant or even suggestive association in the single marker tests. All TDT results for the rs16888748|rs3804320 haplotypes are listed in Appendix4, Table16.  80  4.4.4 Discussion The results from my association analysis of 6p22.2 are consistent with the evidence from both the current NPL analysis (section 3.3) and our past linkage studies, revealing that genetic variation in these two gene clusters are not of major importance in the development of dyslexia in our sample of Canadian families. No variant showed biased transmission in the overall sample for single marker tests. Only one 2-allele haplotype showed a significant p-value in the 2-marker haplotype test, but this result can easily be attributed to the chance of finding a false positive association in 73 tests. Statistically, at the 0.05 level of significance one expects 3.65 false positive associations in 73 applied tests. These two markers, however, do fall within a genomic interval of ~140kb within DCDC2 that may harbour risk variant(s) for dyslexia susceptibility (Deffenbacher et al. 2004; Schumacher et al. 2006).  Only 6.95kb separates the two SNPs involved in the associated haplotypes; rs3804320, the more proximal SNP at 24.285902Mb, is only 0.38kb away from a dyslexia-associated SNP reported by Deffenbacher et al. (2004) and 8.24kb away from a dyslexia-associated SNP reported by Schumacher et al. (2006). It is possible that the single marker association tests in our sample of 83 trios lacked adequate power to detect an association for either of these two markers even though a signal was detected using haplotypes. Alternatively, the DCDC2 gene may have only a small undetectable effect on the dyslexia phenotype in our families. Epistasis (where the phenotypic expression of one gene is modified by one or several other genes) can have a major effect on the phenotypic expression of genetically complex traits, to the extent that effects of interacting loci may not even be detectable when loci are tested individually (Carlborg et al. 2004, 2005, 2006; Bell et al. 2006; Wiltshire et al. 2006). Two genes, DCDC2 (Schumacher et al. 2006, Meng et al. 2005, Deffenbacher et al. 2004) and KIAA0319 (Francks et al. 2004, Cope et al. 2005, Paracchini et al 2006, Harold et al 2006) are rival candidates for the dyslexia susceptibility gene called DYX2. Only 185kb separates the distal cluster VMP/DCDC2/KAAG1 from the more proximal KIAA0319/TTRAP/THEM2 cluster. One study has suggested a possible interaction between KIAA0319 and DCDC2 (Harold et al 2006) while others have found significant association with markers in DCDC2 and failed to detect any association in the proximal cluster (Meng et al 2005; Schumacher et al 2006), or else  81  reported the opposite result, with strong association for KIAA0319 markers and no association for the distal gene cluster (Franks et al 2004; Cope et al 2005; Paracchini et al 2006). Postmortem findings have revealed abnormal neuronal migration and maturation in the brains of individuals with dyslexia (Heir et al 1978; Hynd et al 1983; Galaburda et al 1985; Livingstone 1991). Down-regulation of DCDC2 through RNA interference altered neuronal migration in the developing rat neocortex (Meng et al. 2005), while KIAA0319 was shown to be involved in radial neuronal migration (Paracchini et al 2006). DCDC2 contains two doublecortin homology domains. In vivo and in vitro assays show that doublecortin binds to the microtubule cytoskeleton and doublecortin (DCDC1 gene) has been implicated in X-linked lissencephaly, a severe brain malformation affecting males (Gleeson et al 1998).  DCDC2 is expressed in fetal brain and the adult central nervous system (CNS), and reverse transcriptase polymerase chain reaction (RT-PCR) studies demonstrated that DCDC2 localized to the same regions of the brain responsible for fluent reading, including the inferior and medial temporal cortex, hypothalamus, amygdala and hippocampus (Meng et al 2005).  KIAA0319 is a putative glycosylated single-pass membrane protein.  This protein is also expressed in the fetal brain and adult CNS, including the superior parietal cortex, primary visual cortex and occipital cortex (Meng et al 2005). KIAA0319 displays specific spatial-temporal expression during brain development (Paracchini et al 2006), a finding that is especially pertinent for developmental dyslexia. It is clear that dyslexia is a complex disorder with more than one susceptibility gene. The definition of dyslexia (and accordingly the phenotypic criteria and ascertainment schemes) differs between studies. Consequently, independent studies may be measuring diverse sub- samples of the dyslexia phenotype. The participants in these DYX2 genetic studies have been tested for various phenotypes including phonological awareness, phonological decoding, both timed and untimed single word reading, orthographic decoding, and orthographic choice. Some authors have speculated that the DCDC2 gene may have more relevancy for spelling capacity while the proximal gene cluster including KIAA0319 may be more relevant for word reading ability (Schumacher et al. 2006) This hypothesis is discredited by the detection of association with DCDC2 markers in a sample of individuals with dyslexia who were selected on the basis of word-reading ability by Deffenbacher et al. 2004.  82  A dyslexia susceptibility gene most likely exists in the 6p22 region and there is compelling biological evidence supporting both DCDC2 and KIAA0319 as candidate genes with potential roles in neocortical migration. The conflicting results from independent genetic mapping studies, however, suggest sample-specific genetic heterogeneity within the DYX2 region - some samples might not carry the risk variant(s) at an appreciable allelic frequency. Alternatively, this discrepancy could be the result of random chance operating on several genes of small effect or from difference between studies in the ascertainment scheme and what is being measured as dyslexia. Either (or both) of these two candidate genes may have a major effect on the dyslexia trait variance in the samples that led to their identification, but they do not seem to have a detectable genetic effect in our sample of Canadian families. 4.5 Correcting for multiple comparisons The significance values reported in chapter 4 are pointwise p values and have not been corrected for the number of markers tested for evidence of linkage disequilibrium with a dyslexia locus. The Bonferroni correction, however, is too conservative to account for the multiple testing in the aforementioned association analyses since the markers, correlated due to their LD, are not independent tests. A simple regional assessment of the multiple tests displays a trend for each of the regions linked to PCD in the NPL study to have slightly more than expected significant p values (p<0.05). In the 16p12 region, 10 significant p values were observed out of 176 tests, compared to 8.8 expected significant results (176x0.05= 8.8).  For the 2p21 region, there were 10 observed significant results out of 185 tests (9.25 expected), and for the 4q, there were 13 significant results out of 224 tests (11.2 expected). Thus, in all three regions, the number of markers with p values greater than 0.05 exceeded the number expected. Although the difference is not great, there may not be more than one ‘true positive’ associated (at p=0.05) SNP per region given the sample size of 83 trios. As further support, zero of the 23 tested markers showed association with the DYX2 locus, which did not show linkage in our families. This is less than the 1.15 significant p values that are expected (23x0.05). Therefore, while most detected associations within a dyslexia-linked region are type 1 errors, one or two of the associations, per region, is probably a real signal.  83  Table 6 - Results from single marker association and Haplotype TDT association for 16p12.  Markers with p<0.05 for the AFBAC and TDT are highlighted in blue. P-values >0.05 are displayed in red. * Lowest p value reported for this marker using a TDT analysis of 2-SNP haplotypes, generated by sliding window. Reported p value is for a haplotype between the listed marker and one of the two adjacent markers from the set of 176 markers. See Appendix1, Table13 for a list of 2-SNP haplotype TDT p values for all markers that were significant in single marker association tests.           Single SNP Association 2 SNP Haplotype TDT Association NPL           AFBAC  TDT  Chromosome  Marker  cM  BP  P value  P value  lowest P value *  lod  P value 16p12.3  rs9934960  37.505  17,161,049 0.0313  0.0390  0.0390  NA  NA  16p12.3  rs7202362  40.315  19,025,136 0.0094  0.0126  0.0211  NA  NA  16p12.3  rs4782209  40.361  19,055,889 0.0331  0.0594  0.0211  2.052 0.0011 16p12.3  rs1023442  40.638  19,239,066 0.0457  0.0489  0.0120  1.755 0.0022 16p12.3  rs12447418  40.953  19,447,359 0.0112  0.0056  0.0044  NA  NA  16p12.3  rs12927277  40.960  19,451,482 0.0097  0.0079  0.0044  NA  NA  16p12.3  rs2608173  41.731  19,961,860 0.0256  0.0233  0.0124  1.823 0.0019 16p12.1  rs9939346  45.119  22,658,216 0.0342  0.0285  0.0048  2.136 0.0009 16p12.1  rs10521122  45.135  22,672,084 0.0110  0.0136  0.0048  2.12  0.0009 16p12.1  rs9940479  46.800  23,871,505 0.0160  0.0137  0.0015  1.132 0.0112 16p12.1  rs120908  46.800  23,871,593 0.0023  0.0032  0.0015  NA  NA   84  Table 7 - (7A) Genes found in region chr16:17,161,049-23,871,593 bp. (7B) Copy Number Polymorphisms (CNP) found in region chr16: 17,161,049-23,871,593 bp.  (7A) - Candidate genes discussed in the text are in bold font. Gene position and expression data according to UCSC genome browser on Human Mar. 2006 Assembly.  Gene Symbol  Gene Name  Location (Mb)  Expression  XYLT1  xylosyltransferase I  17.103682‐17.472239 ubiquitous  NPIP  nuclear pore complex interacting protein  18.319299‐18.325514 ubiquitous  NPIP  nuclear pore complex interacting protein  18.359452‐18.378197 ubiquitous  NOMO2  nodal modulator 2 isoform 2  18.418683‐18.480935 ubiquitous  RPS15A  ribosomal protein S15a  18.701778‐18.709157 predominantly non‐brain ARL6IP1  ADP‐ribosylation factor‐like 6 interacting  18.710492‐18.720358 predominantly brain   SMG1  PI‐3‐kinase‐related kinase SMG‐1  18.723676‐18.844887 ubiquitous  TMC7  transmembrane channel‐like 7  18.902831‐18.982002 predominantly non‐brain COQ7  COQ7 protein  18.986428‐18.998853 ubiquitous  LOC162073  hypothetical protein LOC162073  19.032783‐19.040453 ubiquitous  SYT17  B/K protein  19.087139‐19.186055 predominantly brain   TMC5  transmembrane channel‐like 5 isoform a (and isoform b and c)  19.329558‐19.417935 predominantly non‐brain GDE1  Glycerophosphodiester phosphodiesterase 1  19.420516‐19.440951 ubiquitous  CP110  Centrosomal protein of 110 kDa  19.442709‐19.472229 ubiquitous  C16orf62  hypothetical protein LOC57020  19.474541‐19.619945 ubiquitous  IQCK  IQ motif containing K  19.635279‐19.776360 ubiquitous  GPRC5B  G protein‐coupled receptor family C group 5   19.777794‐19.803652 predominantly brain   GPR139  G protein‐coupled receptor 139  19.950544‐19.992601 predominantly brain   GP2  zymogen granule membrane glycoprotein 2 isoform  20.229312‐20.246336 predominantly non‐brain UMOD  uromodulin precursor  20.251874‐20.271538 predominantly non‐brain PDILT  protein disulfide isomerase‐like protein of the  20.277993‐20.323534 ubiquitous           85  Table (7A) - continued  Gene Symbol  Gene Name  Location (Mb)  Expression  ACSM5  acyl‐CoA synthetase medium‐chain family member  20.328357‐20.359782 predominantly non‐brain ACSM2A  acyl‐CoA synthetase medium‐chain family member  20.370360‐20.406492 predominantly non‐brain ACSM2B  acyl‐CoA synthetase medium‐chain family member  20.455584‐20.495196 predominantly non‐brain ACSM1  acyl‐CoA synthetase medium‐chain family member  20.542060‐20.610079 predominantly non‐brain THUMPD1  THUMP domain containing 1  20.652490‐20.660700 predominantly non‐brain ACSM3  SA hypertension‐associated homolog isoform 1 (and isoform 2) 20.682813‐20.715980 predominantly non‐brain LOC81691  exonuclease NEF‐sp  20.725334‐20.768488 predominantly non‐brain DCUN1D3  DCN1 defective in cullin neddylation 1 domain  20.776897‐20.819161 predominantly non‐brain LYRM1  LYR motif containing 1  20.819927‐20.843834 predominantly non‐brain DNAH3  dynein axonemal heavy chain 3  20.851977‐21.078263 predominantly non‐brain TMEM159  transmembrane protein 159  21.077501‐21.099431 predominantly non‐brain ZP2  zona pellucida glycoprotein 2 preproprotein  21.116274‐21.130369 predominantly non‐brain ANKS4B  harmonin‐interacting ankyrin‐repeat containing  21.152517‐21.171251 predominantly non‐brain CRYM  crystallin mu isoform 2 (and isoform 1)  21.177343‐21.221918 predominantly brain   FLJ41766  hypothetical protein LOC400508  21.219671‐21.237413 unknown  RUNDC2B  RUN domain containing 2B  21.298814‐21.310915 unknown  METTL9  methyltransferase like 9 isoform 2 (and isoform 1)  21.518357‐21.576293 predominantly non‐brain IGSF6  immunoglobulin superfamily member 6  21.560107‐21.571473 predominantly non‐brain OTOA  otoancorin isoform 1 (and isoform 2)  21.597336‐21.679551 unknown  KIAA0220  (KIAA0220, UCSC gene prediction)  21.754364‐21.755743 ubiquitous  UQCRC2  ubiquinol‐cytochrome c reductase core protein  21.872110‐21.902169 predominantly non‐brain VWA3A  von Willebrand factor A domain containing 3A  22.011363‐22.075788 unknown  EEF2K  elongation factor‐2 kinase  22.125093‐22.207567 predominantly non‐brain POLR3E  polymerase (RNA) III (DNA directed) polypeptide  22.216242‐22.252844 predominantly non‐brain CDR2  cerebellar degeneration‐related protein 2  22.264758‐22.293439 ubiquitous   KIAA0220  (KIAA0220, UCSC gene prediction)  22.410563‐22.453370 ubiquitous           86  Table (7A) - continued  Gene Symbol  Gene Name  Location (Mb)  Expression  LOC23117  hypothetical protein LOC23117, KIAA0220‐like protein  22.432385‐22.455362 unknown  HS3ST2  heparan sulfate D‐glucosaminyl  22.733361‐22.835160 predominantly brain   USP31  ubiquitin specific peptidase 31  22.980229‐23.068092 ubiquitous  SCNN1G  sodium channel nonvoltage‐gated 1 gamma  23.101541‐23.135701 predominantly non‐brain SCNN1B  sodium channel nonvoltage‐gated 1 beta  23.221092‐23.300121 predominantly non‐brain COG7  component of oligomeric golgi complex 7  23.307317‐23.372004 predominantly non‐brain GGA2  ADP‐ribosylation factor binding protein 2  23.383144‐23.429309 ubiquitous  EARS2  glutamyl‐tRNA synthetase 2  23.440835‐23.476197 ubiquitous  UBFD1  ubiquitin‐binding protein homolog  23.476363‐23.493211 ubiquitous  NDUFAB1  NADH dehydrogenase (ubiquinone) 1 alpha/beta  23.499836‐23.515140 ubiquitous  PALB2  partner and localizer of BRCA2  23.521984‐23.560179 predominantly non‐brain DCTN5  dynactin 5  23.560308‐23.588683 ubiquitous  PLK1  polo‐like kinase  23.597702‐23.609189 predominantly non‐brain ERN2  endoplasmic reticulum to nucleus signalling 2  23.609127‐23.632322 ubiquitous  CHP2  hepatocellular carcinoma antigen gene 520  23.673449‐23.677757 predominantly non‐brain PRKCB1  protein kinase C beta isoform 1 (and isoform 2)  23.754823‐24.139063 predominantly brain      (7B) - Copy Number Polymorphisms (CNP) found in region chr16: 17,161,049-23,871,593 bp. CNP data from SNP and BAC microarrays (Redon) according to UCSC genome browser on Human Mar. 2006 Assembly  CNP #  Position (Mb)  Genomic Size (bp) Band  Variation  cnp1171  17.441602‐19.066807  1625206 16p12.3  gain or loss (unknown direction) cnp1172  20.352825‐20.506848  154024 16p12.3 ‐ 16p12.2 gain or loss (unknown direction) cnp1173  21.212773‐21.856623  643851 16p12.2 ‐ 16p12.1 gain or loss (unknown direction) cnp1174  22.201626‐22.679810  478185 16p12.1  gain or loss (unknown direction)  87  Table 8 - Results from single marker association and Haplotype TDT association for 2p21.  Markers with p<0.05 for the AFBAC and TDT are highlighted in blue. P values >0.05 are displayed in red. * Lowest p value reported for this marker using a TDT analysis of 2-SNP haplotypes, generated by sliding window. Reported p value is for a haplotype between the listed marker and one of the two adjacent markers from the set of 185 markers. See Appendix 2, Table14 for a list of 2-SNP haplotype TDT p values for all markers that were significant in single marker association tests.           Single SNP Association 2 SNP Haplotype TDT Association NPL          AFBAC  TDT  Chromosome  Marker  cM  BP   P value  P value  lowest P value *  lod  P value 2p21  rs2004578  67.349  43,000,362 0.0586  0.0497  0.0364  NA  NA  2p21  rs4953699  67.450  43,052,891 0.0549  0.0466  0.0466  NA  NA  2p21  rs10182679  67.551  43,105,656 0.0190  0.0167  0.0138  1.580 0.0035 2p21  rs13412979  68.433  43,712,594 0.0250  0.0417  0.0085  1.652 0.0029 2p21  rs17335631  68.441  43,727,009 0.0070  0.0076  0.0019  NA  NA  2p21  rs6748247  68.462  43,766,724 0.0111  0.0164  0.0019  1.661 0.0028 2p21  rs11898901  68.884  44,033,240 0.0211  0.0196  0.0196  NA  NA  2p21  rs7583940  68.972  44,109,073 0.0252  0.0140  0.0211  1.808 0.0020 2p21  rs7576649  70.791  45,670,073 0.0215  0.0322  0.0041  1.601 0.0033 2p21  rs4557033  71.268  45,807,828 0.0179  0.0218  0.0296  1.483 0.0045 2p21  rs10865208  71.610  45,910,006 0.0570  0.0407  0.0045  1.603 0.0033 2p21  rs4953273  71.651  45,936,374 0.0057  0.0080  0.0059  1.571 0.0036 2p21  rs10189339  71.669  45,947,264 0.0480  0.0436  0.0059  1.557 0.0037 2p21  rs13036100  71.771  46,012,194 0.0613  0.0455  0.0148  NA  NA   88  Table 9 - Genes found in region chr2:43,105,656-45,947,264 bp.  Candidate genes discussed in the text are in bold font. Gene position and expression data according to UCSC genome browser on Human Mar. 2006 Assembly.   Gene Symbol  Gene Name  Location (Mb)  Expression  ZFP36L2   butyrate response factor 2  43.304210‐43.307249 predominantly non‐brain GITA/3p fusion  death receptor interacting/3p fusion protein splice variant  43.423652‐43.676681 predominantly non‐brain THADA  thyroid adenoma associated isoform 1 (and isoform 2)  43.311479‐43.676689 predominantly non‐brain LOC728819  C1GALT1‐specific chaperone 1‐like  43.755796‐43.756965 unknown  PLEKHH2  pleckstrin homology domain containing family H  43.717951‐43.848630 predominantly non‐brain DYNC2LI1  dynein 2 light intermediate chain isoform 1 (and isoforms 2 and 3) 43.854686‐43.890653 ubiquitous  ABCG5  sterolin 1  43.893115‐43.919462 predominantly non‐brain ABCG8  sterolin 2  43.919607‐43.959109 predominantly non‐brain LRPPRC  leucine‐rich PPR motif‐containing protein  43.968391‐44.076648 ubiquitous  PPM1B  protein phosphatase 1B isoform 4  44.249504‐44.290075 ubiquitous  SLC3A1  solute carrier family 3 member 1  44.356103‐44.401443 predominantly non‐brain PREPL  prolyl endopeptidase‐like isoform C  44.399406‐44.440393 predominantly brain  SIX3  sine oculis homeobox homolog 3  45.022541‐45.025894 ubiquitous  SRBD1  S1 RNA binding domain 1  45.469324‐45.691907 predominantly non‐brain PRKCE  protein kinase C epsilon  45.732547‐46.268633 ubiquitous   89  Table 10 - Results from single marker association and Haplotype TDT association in region 4q13-4q12.  Markers with p<0.05 for the AFBAC and TDT are highlighted in blue. P values >0.05 are displayed in red. * Lowest p value reported for this marker using a TDT analysis of 2-SNP haplotypes, generated by sliding window. Reported p value is for a haplotype between the listed marker and one of the two adjacent markers from the set of 224 markers. See Appendix3, Table15 for a list of 2-SNP haplotype TDT p values for all markers that were significant in single marker association tests.          Single SNP Association 2 SNP Haplotype TDT Association NPL          AFBAC  TDT  Chromosome  Marker  cM  BP  P value  P value  lowest P value *  lod  P value 4q12  rs4864596  72.357  57,291,670 0.0182  0.0282  0.0672  1.400 0.0056 4q12  rs17425986  73.098  57,795,328 0.0984  0.0495  0.0445  1.440 0.0050 4q12  rs4449471  73.446  58,170,248 0.0369  0.0153  0.0050  NA  NA  4q12  rs6838712  73.448  58,171,994 0.0090  0.0071  0.0050  NA  NA  4q12  rs4865252  73.451  58,176,174 0.0135  0.0124  0.0071  NA  NA  4q12  rs1509541  73.605  58,363,698 0.0273  0.0339  0.0241  NA  NA  4q13  rs1397814  74.653  59,961,204 0.0314  0.0348  0.0034  1.805 0.0020 4q13  rs1513537  74.717  60,084,829 0.0212  0.0282  0.0282  NA  NA  4q13  rs1511104  75.174  60,969,781 0.0364  0.0243  0.0031  NA  NA  4q13  rs7690173  75.205  61,030,129 0.0058  0.0061  0.0122  NA  NA  4q13  rs10021608  75.208  61,035,124 0.0479  0.0348  0.0222  NA  NA  4q13  rs13141378  75.212  61,042,489 0.0404  0.0474  0.0254  NA  NA  4q13  rs10033568  75.302  61,217,931 0.0715  0.0396  0.0396  1.557 0.0037 4q13  rs10517524  75.533  61,664,691 0.0071  0.0065  0.0045  1.407 0.0055 4q13  rs11131325  75.667  61,923,342 0.0499  0.0652  0.0628  NA  NA  4q13  rs17238975  75.685  61,950,748 0.0380  0.0364  0.0152  1.169 0.0102   90  Table 11 - (11A) Genes found in region chr4:57,291,670-61,950,748 bp. (11B) Copy number polymorphisms (CNP) found in region chr4:57,291,670-61,950,748 bp.  (11A) - Candidate genes discussed in the text are in bold font. Gene position and expression data according to UCSC genome browser on Human Mar. 2006 Assembly. * Shorter sequence for LPHN3 starts at 62,045,434bp and ends at same position as longer sequence.  Gene Symbol  Gene Name  Location (Mb)  Expression  SPINK2  serine protease inhibitor Kazal type 2  57.370791‐57.382650  predominantly non‐brain REST  RE1‐silencing transcription factor  57.468799‐57.493097  predominantly non‐brain C4orf14  hypothetical protein LOC84273  57.524273‐57.538583  predominantly non‐brain POLR2B  DNA directed RNA polymerase II polypeptide B 57.539866‐57.592091  ubiquitous  IGFBP7  insulin‐like growth factor binding protein 7  57.592001‐57.671296  predominantly non‐brain LPHN3  latrophilin 3 precursor  * 61.749970‐62.620763   predominantly brain      (11B) - CNP data from SNP and BAC microarrays (Redon) according to UCSC genome browser on Human Mar. 2006 Assembly.  CNP #  Position (Mb)  Genomic Size (bp) Band  Variation  cnp 343  chr4:57.748,175‐58.349,780  601,606  4q12  gain or loss (unknown direction) cnp 344  chr4:60.813,754‐60.999,844  186,091  4q13.1 gain or loss (unknown direction) cnp 345  chr4:61.477,441‐61.650,751  173,311  4q13.1 gain or loss (unknown direction)  91  Table 12 - Results from single marker association and Haplotype TDT association for 6p22.2. P values >0.05 are displayed in red. *p value for rs16888748|rs3804320 haplotype  Single SNP Association 2 SNP Haplotype TDT Association NPL             AFBAC  TDT  Chromosome  Gene  Marker  cM  BP  P value  P value  * P value<0.05  lod  P  value 6p22.2     rs9379635  47.541 24,175,527 0.8459  0.8415  ‐  ‐0.008 0.5779  6p22.2     rs10806983  47.586 24,202,874 1.0000  1.0000  ‐  NA  NA  6p22.2  VMP  rs7455023  47.649 24,241,681 0.5828  0.5831  ‐  ‐0.017 0.6101  6p22.2     rs13191953  47.677 24,259,022 0.3452  0.3692  ‐  NA  NA  6p22.2     rs16888748  47.709 24,278,954 0.7584  0.6310  0.0196  NA  NA  6p22.2  DCDC2  rs3804320  47.721 24,285,902 0.3915  0.3865  0.0196  NA  NA  6p22.2  DCDC2  rs1770909  47.793 24,330,431 0.8701  0.8658  ‐  ‐0.035 0.6565  6p22.2  DCDC2  rs9460984  47.833 24,355,227 0.8674  0.8694  ‐  ‐0.038 0.6616  6p22.2  DCDC2  rs9295618  47.836 24,356,599 0.6522  0.7290  ‐  NA  NA  6p22.2  DCDC2  rs17302512  47.889 24,389,314 0.8702  0.8728  ‐  NA  NA  6p22.2  DCDC2  rs793698  48.052 24,485,324 0.1361  0.1129  ‐  ‐0.043 0.6712  6p22.2     rs2753922  48.122 24,525,645 0.7282  0.7324  ‐  NA  NA  6p22.2     rs1056285  48.127 24,528,523 1.0000  0.8527  ‐  ‐0.043 0.6726  6p22.2     rs2744576  48.197 24,568,686 0.8236  0.9093  ‐  ‐0.044 0.6740  6p22.2     rs9467176  48.204 24,572,886 0.5090  0.5164  ‐  ‐0.044 0.6744  6p22.2     rs2760139  48.232 24,589,278 0.2723  0.2513  ‐  NA  NA  6p22.2     rs2744597  48.322 24,641,053 0.4570  0.4652  ‐  ‐0.048 0.6809  6p22.2  KIAA0319  rs2744605  48.342 24,652,311 0.8033  0.7963  ‐  NA  NA  6p22.2  KIAA0319  rs2817250  48.357 24,661,360 0.5168  0.5078  ‐  ‐0.049 0.6828  6p22.2  KIAA0319  rs9467239  48.480 24,732,836 0.2078  0.2164  ‐  NA  NA  6p22.2  KIAA0319  rs761101  48.494 24,740,511 0.1010  0.0990  ‐  NA  NA  6p22.2  TTRAP  rs3756819  48.551 24,773,319 1.0000  1.0000  ‐  ‐0.038 0.6627  6p22.2     rs1980643  48.723 24,872,767 0.6239  0.6219  ‐  ‐0.035 0.6560   92  CHAPTER 5 CONCLUSION The aim of this study was to locate and identify genes predisposing to phonological coding dyslexia (PCD). A genome-wide linkage study and four targeted fine-mapping family-based association studies were performed to screen for dyslexia susceptibility genes in 101 Canadian families with multiple PCD-affected individuals. The most significant finding was the identification of suggestive evidence for linkage with PCD at two novel regions, 16p12 and 4q12-q13. This result was obtained using a model-free (non-parametric) linkage analysis yielding lod scores of 2.263 and 1.879, respectively, for the two regions. While this study is the first to report linkage for 16p12 and 4q12-q13 to dyslexia, both regions have shown linkage to bipolar disorder, and the former has also been linked to ADHD and autism. The NPL analysis also provided independent confirmation of linkage to the well-replicated DYX3 locus (at 2p21) with a lod of 1.831 but failed to detect linkage to DYX2 (6p22.2). This analysis replicates a previous report of no linkage in our families to DYX2, generated with more informative but less numerous microsatellite markers. The genetic analyses in this thesis were carried out with a high-density set of SNP markers, which allowed a finer localization of the linkage signals but also introduced new obstacles. Not only do current genetic-linkage software programs require markers to be in linkage equilibrium, but it remains computationally infeasible to perform multipoint linkage analysis with >50K markers in larger multigenerational pedigrees. Four regions (16p12, 2p21, 4q12-q13 and 6p22.2) were tested for association with PCD in 83 trios, a subset of the 101 families, using the TDT and the AFBAC test.  Association was detected in each of the three regions that were linked to PCD in the NPL scan, while none of the tested markers was associated with PCD in the 6p22.2 region. My results confirm that the putative dyslexia-candidate genes at 6p22.2 do not have a detectable genetic effect in our sample of Canadian families. The sample size of 83 trios gave limited ability to detect a significant association (where the p value could withstand correction for the problem of multiple testing with a number of markers). Demonstrating appropriate statistical significance is a challenge in the genetic analysis of complex traits. The methods of correction employed for the problem of  93  multiple testing reduce the chance of reporting a false positive but also compromise the detection of true positives, especially with complex traits that display incomplete penetrance, polygenic inheritance, and genetic heterogeneity. Four candidate genes (two at 16p12 and one each at 2p21 and 4q12) were identified by markers associated with PCD as possible candidates for a genetic predisposition to dyslexia. Two of these candidates belong to the same gene family of protein kinases: PRKCB1 (at 16p12) and PRKCE (at 2p21). The sample size in these association studies was smaller than ideally required and the results, therefore, are not definitive but instead relevant as part of an ongoing study of the genetics of dyslexia susceptibility. The novel 16p12 dyslexia-linked region is a valuable area to pursue in future work based off the analyses in this thesis. The complex genetic architecture of the 16p12 region should be studied through further linkage and association analyses using copy number variants (CNVs) as markers, since the raw data collected from the Affymetrix SNP arrays can also be used for CNV analysis. The four possible candidate genes are also prime targets for genetic interactions studies. Association studies identify alleles that are statistically correlated with a clinically relevant phenotype in affected individuals. Genetic association does not prove causality, nor can it distinguish between a functional variant and other neighbouring SNPs that are in strong linkage disequilibrium. Once a sequence variant has been identified, mutational and functional studies, demonstrating that the variant alters the function or regulation of a candidate gene in individuals with dyslexia, are required as definitive evidence that the variant is indeed a dyslexia risk allele. Other approaches, such as animal model systems or comparing expression levels in DNA microarray screens of genes in the central nervous system, can also provide indirect evidence that a candidate gene plays an aetiological role in dyslexia susceptibility. Genetic mapping is the first step in isolating genes involved in dyslexia predisposition, and determining the actual functional variants that are responsible for increased susceptibility to dyslexia is important for diagnosis and therapy.  Reading impediments can be overcome in some cases with early educational intervention and diverse compensatory strategies.  With the identification of dyslexia risk alleles, we may be able to develop a genetic profile that targets an individual for specific behavioural interventions and therapies.  Further characterization of  94  dyslexia-susceptibility genes and their role in neurodevelopment may enable us to decipher the mechanisms that underlie reading, language impairment and cognitive skills.   95  REFERENCES  Abecasis GR, Cherny SS, Cookson WO, Cardon LR (2002) Merlin-rapid analysis of dense genetic maps using sparse gene flow trees. Nat Genet 30:97-101.  Abreu, PC, Greenberg, DA, and Hodge, SE (1999) Direct power comparisons between simple LOD scores and NPL scores for linkage analysis in complex diseases. Am J Hum Genet 65: 847 857.  Anderson JS et al (2003) Proteomic characterization of the human centrosome by protein correlation profiling. Nature 426:570-574.  Andrews W, Liapi A, Plachez C, Camurri L, Zhang J, Mori S, Murakami F, Parnavelas JG, Sundaresan V, Richards LJ (2006) Robo1 regulates the development of major axon tracts and interneuron migration in the forebrain. Development 133:2243-2252.  Ansley SJ, Badano JL, Blacque OE, Hill J, Hoskins BE, Leitch CC, Kim JC, Ross AJ, Eichers ER, Teslovich TM, Mah AK, Johnsen RC, Cavender JC, Lewis RA, Leroux MR, Beales PL, Katsanis N (2003) Basal body dysfunction is a likely cause of pleiotropic Bardet-Biedl syndrome. Nature 425:628-633.  Anthoni H, Zucchelli M, Matsson H, Müller-Myhsok B, Fransson I, Schumacher J, Massinen S, Onkamo P, Warnke A, Griesemann H, Hoffmann P, Nopola-Hemmi J, Lyytinen H, Schulte- Körne G, Kere J, Nöthen MM, Peyrard-Janvid M (2007) A locus on 2p12 containing the co- regulated MRPL19 and C2ORF3 genes is associated to dyslexia. Hum Mol Genet 16:667-677.  August GJ, Garfinkel BD (1989) Behavioral and cognitive subtypes of ADHD. J Am Acad Child Adolesc Psychiatry 28:739-748.  August GJ, Garfinkel BD (1990) Comorbidity of ADHD and reading disability among clinic- referred children. J Abnorm Child Psychol 18:29-45.  Badano JL, Teslovich TM, Katsanis N (2005) The centrosome in human genetic disease. Nature 6: 194-207.  Bakwin H (1973) Reading disability in twins. Develop Med Child Neurol 15:184-187.  Barnby G, Abbott A, Sykes N, Morris A, Weeks DE, Mott R, Lamb J, Bailey AJ, Monaco AP; International Molecular Genetics Study of Autism Consortium (2005) Candidate-gene screening and association analysis at the autism-susceptibility locus on chromosome 16p: evidence of association at GRIN2A and ABAT. Am J Hum Genet 76:950-966.  Barrett JC, Fry B, Maller J, Daly MJ (2005) Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 21:263-265.  96   Bates TC, Luciano M, Castles A, Coltheart M, Wright MJ, Martin NG (2007) Replication of reported linkages for dyslexia and spelling and suggestive evidence for novel regions on chromosomes 4 and 17. Eur J Hum Genet15:194-203.  Bell J Tzenova, Kaplan BJ, Petryshen TL, Hsiung R, Field LL. Evidence for gene-interactions among established dyslexia susceptibility loci. (manuscript in preparation).  Bell J Tzenova, Wallace C, Dobson R, Wiltshire S, Mein C, Pembroke J, Brown M, Clayton D, Samani N, Dominiczak A, Webster J, Lathrop GM, Connell J, Munroe P, Caulfield M, Farrall M (2006) Two-dimensional genome-scan identifies novel epistatic loci for essential hypertension. Hum Mol Genet 15:1365-74  Ben-Ari Y, Aniksztejn L, Bregestovski P (1992) Protein kinase C modulation of NMDA currents: an important link for LTP induction. Treds Neurosci 15:333-339.  Bisgaard ML, Eiberg H, Moller N, Niebuhr E, Mohr J (1987) Dyslexia and chromosome 15 heteromorphism: negative lod score in Danish material. Clin Genet 32:118-119.  Bin Sun H, Ruan Y, Xu ZC, Yokota H (2002) Involvement of the calcium-independent receptor for alpha-latrotoxin in brain ischemia.  Mol Brain Res 104: 246-249.  Borjigin J, Deng J, Sun X, De Jesus M, Liu T, Wang MM. (2003) Diurnal pineal 3-O- sulphotransferase 2 expression controlled by beta-adrenergic repression. J Biol Chem 278:16315-16319.  Brkanac Z, Chapman NH, Matsushita MM, Chun L, Nielsen K, Cochrane E, Berninger VW, Wijsman EM, Raskind WH (2007) Evaluation of candidate genes for DYX1 and DYX2 in families with dyslexia. Am J Med Genet B Neuropsychiatr Genet 144:556-560.  Breitmeyer BG (1993) The roles of the sustained (P) and transient (M) channels in reading and reading disability. In: Wright SF, Groner R (eds) Facets of Dyslexia and its Remediation, North- Holland, Amsterdam, pp 13-31.  BRLMM: an Improved Genotype Calling Method for the GeneChip® Human Mapping 500K Array Set. Revision Date: 2006-04-14 Revision Version: 1.0 http://www.affymetrix.com/support/technical/whitepapers/brlmm_whitepaper.pdf  Brown,WE, Eliez S, Menon V, Rumsey JM, White CD, Reiss AL (2001) Preliminary evidence of widespread morphological variations of the brain in dyslexia. Neurology 56:781-783.  Brunswick N, McCrory E, Price CJ, Frith CD, Frith U (1999) Explicit and implicit processing of words and pseudowords by adult developmental dyslexics: A search for wernicke's wortschatz? Brain 122 (Pt 10):1901-1917.   97  Bult A, Kobylk ME, Van der Zee EA (2001) Differential expression of protein kinase C betaI (PKCbetaI) but not PKCalpha and BKCbetaII in the suprachiasmatic nucleus of selected house mouse lines, and the relationship to arginine-vasopressin. Brain Res 1141:205-213.  Cardon LR, Bell JI (2001) Association study designs for complex diseases. Nat Rev Genet 2:91- 99.  Cardon L, Smith S, Fulker DW, Kimberling WJ, Pennington BF, DeFries JC (1994) Quantitative Trait Locus for Reading Disability on Chromosome 6. Science 266:276-279.  Cardon LR, Smith SD, Fulker DW, Kimberling WJ, Pennington BF, DeFries JC (1995) Quantitative trait locus for reading disability: Correction. Science 268:1553.  Carlborg O, Hocking PM, Burt DW, Haley CS (2004) Simultaneous mapping of epistatic QTL in chickens reveals clusters of QTL pairs with similar genetic effects on growth. Genet Res 83:197- 209.  Carlborg O, Brockmann GA, Haley CS (2005) Simultaneous mapping of epistatic QTL in DU6i x DBA/2 mice. Mamm Genome16:481-94.  Carlborg O, Jacobsson L, Ahgren P, Siegel P, Andersson L (2006) Epistasis and the release of genetic variation during long-term selection. Nat Genet 38:418-20.  Carlson CS, Eberle MA, Rieder MJ, Yi Q, Kruglyak L, Nickerson DA (2004) Selecting a maximally informative set of single-nucleotide polymorphisms for association analyses using linkage disequilibrium. Am J Hum Genet 74:106–120.  Cassidy F, Zhao C, Badger J, Claffey E, Dobrin S, Roche S, McKeon P (2007) Genome-Wide Scan of Bipolar Disorder and Investigation of Population Stratification Effects on Linkage: Support for Susceptibility Loci at 4q21, 7q36, 9p21, 12q24, 14q24, and 16p13. Am J Med Genet Part B 144B:791-801.  Castles A, Coltheart M (1993) Varieties of developmental dyslexia. Cognition 47:149-180.  Chadwick O, Taylor E, Taylor A, Heptinstall E, Danckaerts M (1999) Hyperactivity and reading disability: a longitudinal study of the nature of the association. J Child Psychol Psychiatry 40:1039–1050.  Chapman NH, Raskind WH, Thomson JB, Berninger VW, Wijsman EM (2003) Segregation analysis of phenotypic components of learning disabilities. II. Phonological decoding. Am J Med Genet B Neuropsychiatr Genet 121:60-70.  Chapman NH, Igo RP, Thomson JB, Matsushita M, Brkanac Z, Holzman T, Berninger VW, Wijsman EM, Raskind WH (2004) Linkage analyses of four regions previously implicated in  98  dyslexia: confirmation of a locus on chromosome 15q. Am J Med Genet B Neuropsychiatr Genet 131:67-75.  Chase CH, Rosen GD, Sherman GF (eds) (1996) Developmental dyslexia: neural, cognitive, and genetic mechanisms. York Press, Baltimore, MD.  Chen Z, IndjeianVB, McManus M, Wang L, Dynlacht BD (2002) CP110, a Cell Cycle- Dependent CDK Substrate, Regulates Centrosome Duplication in Human Cells. Dev Cell 3:339- 350.  Clark C, Prior M, Kinsella G (2002) The relationship between executive function abilities, adaptive behaviour, and academic achievement in children with externalising behaviour problems. J Child Psychol Psychiatry 43:785–796.  Colombo PJ, Wetsel WC, Gallagher M (1997) Spatial memory is related to hippocampal subcellular concentrations of calcium-dependent protein kinase C isoforms in young and aged rats. Proc Natl Acad Sci USA 94:14195-14199.  Cope N, Harold D, Hill G, Moskvina V, Stevenson J, Holmans P, Owen MJ, O'Donovan MC, Williams J (2005)  Strong evidence that KIAA0319 on chromosome 6p is a susceptibility gene for developmental dyslexia. Am J Hum Genet 76:581-591.  Cornelissen P, Richardson A, Mason A, Fowler S, Stein (1995) Contrast sensitivity and coherent motion detection measured at photopic luminance levels in dyslexics and controls. Vision Res 35:1483-94.  Davletov BA, Shamotienko OG, Lelianova VG, Grishin EV, Ushkaryov YA (1996) Isolation and biochemical characterization of a Ca2+-independent α-latrotoxin-binding protein. J Biol Chem 271:23239–23245.  Daw EW, Heath SC, Lu Y (2005) Single-nucleotide polymorphism versus microsatellite markers in a combined linkage and segregation analysis of a quantitative trait. BMC Genet; 6(Suppl 1): S32.  de Bakker PIW, Yelensky R, Pe'er I, Gabriel SB, Daly MJ, Altshuler D (2005) Efficiency and power in genetic association studies. Nature Genetics 37:1217-1223.  de Kovel CG, Hol FA, Heister JG, Willemen JJ, Sandkuijl LA, Franke B, Padberg GW (2004) Genomewide scan identifies susceptibility locus for dyslexia on Xq27 in an extended Dutch family. J Med Genet 41:652-657.  de Kovel CGF, Franke B, Hol FA, Lebrec JJP, Maassen B, Brunner H, Padberg GW, Platko J, Pauls D (2008) Confirmation of Dyslexia Susceptibility Loci on Chromosomes 1p and 2p, but Not 6p in a Dutch Sib-Pair Collection. Am J Med Genet Part B 147B: 294-300.   99  Decker LV, Parker PJ (1994) Protein Kinase C: a question of specificity. Trends biochem Sci 19:73-77.  Deffenbacher KE, Kenyon JB, Hoover DM, Olson RK, Pennington BF, DeFries JC, Smith SD (2004) Refinement of the 6p21.3 quantitative trait locus influencing dyslexia: linkage and association analyses. Hum Genet 115:128-138.  DeFries JC, Alarcon M (1996) Genetics of specific reading disability. Ment Retard Dev Disabil Res Rev 2:39-47.  DeFries JC, Fulker D, LaBuda MC (1987) Evidence for a genetic etiology in reading disability of twins. Nature 329:537-539.  DeFries JC (1989) Gender ratios in reading-disabled children and their affected relatives: a commentary. J Learning Disabilities 22:544-545.  de Kovel CG, Hol FA, Heister JG, Willemen JJ, Sandkuijl LA, Franke B, Padberg GW (2004) Genomewide scan identifies susceptibility locus for dyslexia on Xq27 in an extended Dutch family. J Med Genet 41:652-7.  Demb JB, Boynton GM, Best M, Heeger DJ (1998) Psychophysical evidence for a magnocellular pathway deficit in dyslexia. Vision Res 38:1555-9.  Dencla MB, Rudel GR (1976) Rapid ’automatized’ naming (R.A.N.): Dyslexia differentiated from other learning disabilities. Neuropsychol 14:471-479.  Di X, Matsuzaki H, Webster TA, Hubbell E, Liu G, Dong S, Bartell D, Huang J, Chiles R, Yang G, Shen MM, Kulp D, Kennedy GC, Mei R, Jones KW, Cawley S (2005) Dynamic Model Based Algorithms for Screening and Genotyping over 100K SNPs on Oligonucleotide Microarrays. Bioinformatics 21:1958-63.  Duara R, Jushch A, Gross-Glen K, Barker W, Jallad B, Pascal S, Loewenstein DA et al (1991) Neuroanatomic differences between dyslexic and normal readers on magnetic resonance imaging scans. Arch Neurol 48:410-416  Dykman RA, Ackerman PT (1991) Attention deficit disorder and specific reading disability: Separate but often overlapping disorders. J Learn Disabil 24:96-103.  Eden GF, VanMeter JW, Rumsey JM, Maisog JM, Woods RP and Zeffiro TA (1996) Abnormal processing of visual motion dyslexia revealed by functional brain imaging. Nature 382:66-69.  Eichler EE, Johnson ME, Alkan C, Tuzun E, Sahinalp C, Misceo D, Archidiacono N, Rocchi M (2001) Divergent Origins and Concerted Expansion of Two Segmental Duplications on Chromosome 16. J Hered 92:462-468.   100  Fagerheim T, Raeymaekers P, Tonnessen FE, Pedersen M, Tranebjaerg L, Lubs HA (1999) A new gene (DYX3) for dyslexia is located on chromosome 2. J Med Genet 36:664-9.  Fagerheim T (2002) Refinement of the dyslexia 3 locus (DYX3) on chromosome 2p15-16 and mutation screening of candidate genes. Am J Hum Genet 71(S): 461.  Fagerström S, Påhlman S, Gestblom C, Nånber E (1996) Protein Kinase C-ε Is Implicated in Neurite Outgrowth in Differentiating Human Neuroblastoma Cells. Faraone SV, Biederman J, Weiffenbach B, Keith T, Chu MP, Weaver A, Spencer TJ, Wilens TE, Frazier J, Cleves M, Sakai J (1999) Dopamine D4 gene 7-repeat allele and attention deficit hyperactivity disorder. Am J Psychiatry 156:768-770  Faraone SV, Doyle AE, Mick E, Biederman J (2001) Meta-analysis of the association between the 7-repeat allele of the dopamine D(4) receptor gene and attention deficit hyperactivity disorder. Am J Psychiatry  158:1052-1057.  Faraone SV, Perlis RH, Doyle AE, Smoller JW, Goralnick JJ, Holmgren MA, Sklar P (2005) Molecular genetics of attention-deficit/hyperactivity disorder. Biol Psychiatry 57:1313-23. Review. Fawcett AJ, Nicolson RR, Dean P (1996) Impaired performance of children with dyslexia on a range of cerebellar tasks. Ann Dyslexia 46:259-283.  Fawcett AJ, Nicolson RI (1999) Performance of dyslexic children on cerebellar and cognitive tests. J Mot Behav 31:68-78.  Feng J, Schroer R, Yan J, Song W, Yang C, Bockholt A, Cook EH, Skinner C, Schwartz CE, Sommer SS (2006) High frequency of neurexin 1beta signal peptide structural variants in patients with autism. Neurosci Lett 409:10 13.  Field LL, Kaplan B (1998) Absence of linkage of phonological coding dyslexia to chromosome 6p23-p21.3 in a large family dataset. Amer J Hum Genet 63:1448-1456. Erratum 1999 (table omitted by publisher): Amer J Hum Genet 64:334.  Finucci JM, Guthrie JT, Childs AL, Abbey H, Childs B (1976) The genetics of specific reading disability. Ann Hum Genet 40:1-23.  Fisher PJ, Turic D, Williams NM, McGuffin P, Asherson P, Ball D, Craig I, Eley T, Hill L, Chorney K, Chorney MJ, Benbow CP, Lubinski D, Plomin R, Owen MJ (1999) DNA pooling identifies QTLs on chromosome 4 for general cognitive ability in children. Hum Mol Genet 8:915-22.  Fisher SE, Marlow AJ, Lamb J, Maestrini E, Williams DF, Richardson AJ, Weeks DE, Stein JF, Monaco AP (1999) A quantitative-trait locus on chromosome 6p influences different aspects of developmental dyslexia. Am J Hum Genet 64:146-156.   101  Fisher SE, DeFries JC (2002) Developmental dyslexia: Genetic dissection of a complex cognitive trait. Nat Rev Neurosci 3:767-780.   Fisher SE, Francks C, Marlow AJ, MacPhie IL, Newbury DF, Cardon LR, Ishikawa-Brush Y, Richardson AJ, Talcott JB, Gayan J, Olson RK, Pennington BF, Smith SD, DeFries JC, Stein JF, Monaco AP (2002a)  Independent genome-wide scans identify a chromosome 18 quantitative- trait locus influencing dyslexia. Nat Genet 30:86-91.  Fisher SE, Francks C, McCracken JT, McGough JJ, Marlow AJ, MacPhie IL, Newbury DF, Crawford LR, Palmer CG, Woodward JA, Del'Homme M, Cantwell DP, Nelson SF, Monaco AP, Smalley SL (2002b). A genomewide scan for loci involved in attention-deficit/hyperactivity disorder. Am J Hum Genet 70:1183-1196.  Fitch RH, Tallal P (2003) Neural mechanisms of language-based learning impairments: insights from human populations and animal models. Behav Cogn Neurosci Rev 2:155-178.  Francks C, Fisher SE, Olson RK, Pennington BF, Smith SD, DeFries JC, Monaco AP (2002) Fine mapping of the chromosome 2p12-16 dyslexia susceptibility locus: Quantitative association analysis and positional candidate genes SEMA4F and OTX1. Psychiatr Genet 12:35-41.  Francks C, Paracchini S, Smith SD, Richardson AJ, Scerri TS, Cardon LR, Marlow AJ, MacPhie IL, Walter J, Pennington BF, Fisher SE, Olson RK, DeFries JC, Stein JF, Monaco AP (2004) A 77-kilobase region of chromosome 6p22.2 is associated with dyslexia in families from the United Kingdom and from the United States. Am J Hum Genet 75:1046-58.  Frenkel M, Sherman GF, Bashan KA, Galaburda AM, LoTurco JJ (2000) Neocortical ectopias are associated with attenuated neurophysiological responses to rapidly changing auditory stimuli. Neuroreport 11:575-579.  Frith U (1998) Cognitive deficits in developmental disorders. Scan J Psychol 39:191-195.  Froster U, Schulte-Korne G, Hebebrand J, Remschmidt H (1993) Cosegregation of balanced translocation (1;2) with retarded speech development and dyslexia. Lancet 342:178-179.  Galaburda AM, Menard MT, Rosen GD (1994) Evidence for aberrant auditory anatomy in developmental dsylexia. Proc Nat Acad Sci (USA) 91:8010-8013.  Galaburda AM, Sherman GF, Rosen GD, Aboitiz F, Geschwind N (1985) Developmental dyslexia: four consecutive patients with cortical anomalies. Ann Neurology 18:222-233.  Galaburda AM (1989) Ordinary and Extraordinary Brain Development: Anatomical Variation in Developmental Dyslexia.  Annals of Dyslexia 39:67-80.   102  Galaburda AM, LoTurco J, Ramus F, Fitch RH, Rosen GD (2006) From genes to behavior in developmental dyslexia. Nat Neurosci 9:1213-1217.  Gayan J, Olson RK (2003) Genetic and environmental influences on individual differences in printed word recognition. J Exp Child Psychol 84:97-123.  Gayan J, Smith SD, Cherny SS, Cardon LR, Fulker DW, Brower AM, Olson RK, Pennington BF, DeFries JC (1999) Quantitative-trait locus for specific language and reading deficits on chromosome 6p. Am J Hum Genet 64:157-164.  Gayan J, Willcutt EG, Fisher SE, Francks C, Cardon LR, Olson RK, Pennington BF, Smith SD, Monaco AP, DeFries JC (2005) Bivariate linkage scan for reading disability and attention- deficit/hyperactivity disorder localizes pleiotropic loci. J Child Psychol Psychiatry 46:1045- 1056.  Gilger JW, Pennington BF, DeFries JC (1992) A twin study of the etiology of comorbidity: attention-deficit hyperactivity disorder and dyslexia. Am Acad Child Adolesc Psychiatry 31:343- 348.  Gleeson JG, Allen KM, Fox JW, Lamperti ED, Berkovic S, Scheffer I, Cooper EC, Dobyns WB, Minnerath SR, Ross ME, Walsh CA (1998) Doublecortin, a brain-specific gene mutated in human X-linked lissencephaly and double cortex syndrome, encodes a putative signaling protein. Cell 92:63-72.  Goode EL, Badzioch MD, Jarvik GP (2005) Bias of allele-sharing linkage statistics in the presence of intermarker linkage disequilibrium. BMC Genet. 6(Suppl 1): S82.  Gotlieb AI, May LM, Subrahmanyan L, Kalnins VI (1981) Distribution of microtubule organizing centers in migrating sheets of endothelial cells. J Cell Biol 91:589-594.  Grigorenko EL, Wood FB, Meyer MS, Hart LA, Speed WC, Shuster A, Pauls DL (1997) Susceptibility loci for distinct components of developmental dyslexia on chromosomes 6 and 15. Am J Hum Genet 60:27-39.  Grigorenko EL, Wood FB, Meyer MS, Pauls DL (2000) Chromosome 6p influences on different dyslexia-related cognitive processes: further confirmation. Am J Hum Genet 66:715-723.   Grigorenko EL, Wood FB, Meyer MS, Pauls JE, Hart LA, Pauls DL (2001) Linkage studies suggest a possible locus for developmental dyslexia on chromosome 1p. Am J Med Genet 105:120-129.  Grigorenko EL, Wood FB, Golovyan L, Meyer M, Romano C, Pauls D (2003) Continuing the search for dyslexia genes on 6p. Am J Med Genet B Neuropsychiatr Genet 118:89-98.   103  Grigorenko EL (2005) A conservative Meta-Analysis of Linkage and Linkage-Association Studies of Developmental Dyslexia. Sci Studies of Reading 9:285-316.  Habib M (2000) The neurological basis of developmental dyslexia. An overview and working hypothesis. Brain 123:2373-2399.  Hallgren B (1950) Specific dyslexia ("congenital word blindness"): a clinical and genetic study. Acta Psychiatr Scand 65:1-287.  Hannula-Jouppi K, Kaminen-Ahola N, Taipale M, Eklund R, Nopola-Hemmi J, Kaariainen H, Kere J (2005) The axon guidance receptor gene ROBO1 is a candidate gene for developmental dyslexia. PLoS Genet 1(4):e50.  Harold D, Paracchini S, Scerri T, Dennis M, Cope N, Hill G, Moskvina V, Walter J, Richardson AJ, Owen MJ, Stein JF, Green ED, O'Donovan MC, Williams J, Monaco AP (2006) Further evidence that the KIAA0319 gene confers susceptibility to developmental dyslexia. Mol Psychiatry 11:1085-1091, 1061.  Heim S, Keil A (2004)  Large-scale neural correlates of developmental dyslexia. Eur Child Adolesc Psychiatry13:125-40. Review.  Helmuth (2001) Dyslexia: Same Brains, Different Languages. Science 16 March: 2064.  Hier DB, Lemay M, Rosenberger PB, Perlo VP (1978) Developmental dyslexia: evidence for a subgroup with a reversal of cerebral asymmetry. Arch Neurol 35:90-92.  Hinshaw SP (1992) Academic underachievement, attention deficits, and aggression: comorbidity and implications for intervention. J Consult Clin Psychol 60:893–903.  Hodge CW, Raber J, McMahon T, Walter H, Sanchez-Perez AM, Olive MF, Mehmert K, Morrow AL, Messing RO (2002) Decreased anxiety-like behavior, reduced stress hormones, and neurosteroid supersensitivity in mice lacking protein kinase C epsilon. J Clin Invest. 110:1003- 1010.  Holmes J, Payton A, Barrett JH, Hever T, Fitzpatrick H, Trumper AL, Harrington R, McGuffin P, Owen M, Ollier W, Worthington J, Thapar A (2000) A family-based and case-control association study of the dopamine D4 receptor gene and dopamine transporter gene in attention deficit hyperactivity disorder. Mol Psychiatry  5:523-530.  Hsiung G-Y R, Kaplan BJ, Petryshen TL, Lu S, Field LL (2004) A dyslexia susceptibility locus (DYX7) linked to dopamine D4 receptor (DRD4) region on chromosome 11p15.5.  Am J Med Genet (Neuropsych Genet) 125B:112-119.  Hu H (1999) Chemorepulsion of neuronal migration by Slit2 in the developing mammalian forebrain. Neuron 23:703-711.  104   Hugdahl K, Heiervang E, Ersland L, Lundervold A, Steinmetz H, Smievoll AI (2003) Significant relation between MR measures of planum temporale area and dichotic processing of syllables in dyslexic children. Neuropsychologia 41:666-75.  Hynd G, Cohen M (1983) Dyslexia: Neuropsychological Theory, Research, and Clinical Differentiation. Grune and Stratton, New York, NY.  Hynd GW, Semrud-Clikeman M, Lorys AR, Novey ES, Eliopulos D (1990) Brain morphology in developmental dyslexia and attention deficit/hyperactivity. Arch Neurol 47:919-926.  Iafrate AJ, Feuk L, Rivera MN, Listewnik ML, Donahoe PK, Qi Y, Scherer SW, Lee C (2004) Detection of large-scale variation in the human genome. Nature Genet 36:949-951.  Ichtchenko K, Bittner MA, Krasnoperov V, Little AR, Chepurny O, Holz RW, Petrenko AG. (1999) A novel ubiquitously expressed alpha-latrotoxin receptor is a member of the CIRL family of G-protein-coupled receptors. J Biol Chem 274:5491-5498.  Ivaska J, Whelan RDH, Watson R, Parker J (2002) PKC epsilon controls the traffic of beta1 integrins in motile cells. EMBO J 21:3608-3619.  Jastak S, Wilkinson GS (1984) The wide range achievement test - revised. Jastak Associates, Wilmington, DE.  Johnson ME, Viggiano L, Bailey JA, Abdul-Rauf M, Goodwin G, Rocchi M, Eichler EE (2001) Positive selection of a gene family during the emergence of humans and African apes. Nature 413:514-519.  Johnson ME; National Institute of Health Intramural Sequencing Center Comparative Sequencing Program, Cheng Z, Morrison VA, Scherer S, Ventura M, Gibbs RA, Green ED, Eichler EE (2006) Recurrent duplication-driven transposition of DNA during hominoid evolution. Proc Natl Acad Sci USA 103:17626-17631.  Jones HB (1998) The relative power of linkage and association studies for the detection of genes involved in hypertension.  Kidney Int 53:1446-1448.  Kaminen N, Hannula-Jouppi K, Kestila M, Lahermo P, Muller K, Kaaranen M, Myllyluoma B, Voutilainen A, Lyytinen H, Nopola-Hemmi J, Kere J (2003)  A genome scan for developmental dyslexia confirms linkage to chromosome 2p11 and suggests a new locus on 7q32. J Med Genet 40:340-345.  Kaplan DE, Gayan J, Ahn J, Won TW, Pauls D, Olson RK, DeFries JC, Wood F, Pennington BF, Page GP, Smith SD, Gruen JR (2002) Evidence for linkage and association with reading disability on 6p21.3-22. Am J Hum Genet 70:1287-1298.   105  Karolchik D, Kuhn, RM, Baertsch R, Barber GP, Clawson H, Diekhans M, Giardine B, Harte RA, Hinrichs AS, Hsu F, Miller W, Pedersen JS, Pohl A, Raney BJ, Rhead B, Rosenbloom KR, Smith KE, Stanke M, Thakkapallayil A, Trumbower H, Wang T, Zweig AS, Haussler D, Kent WJ (2008) The UCSC Genome Browser Database: 2008 update. Nucleic Acids Res 36:D773- 779.  Kent WJ, Sugnet CW, Furey TS, Roskin KM, Pringle TH, Zahler AM, Haussler D (2002) The human genome browser at UCSC. Genome Res 6:996-1006.  Kidd T, Brose K, Mitchell KJ, Fetter RD, Tessier-Lavigne M, Goodman CS, Tear G (1998) Roundabout controls axon crossing of the CNS midline and defines a novel subfamily of evolutionarily conserved guidance receptors. Cell 92:205-215.  Kim HG, Kishikawa S, Higgins AW, Seong IS, Donovan DJ, Shen Y, Lally E, Weiss LA, Najm J, Kutsche K, Descartes M, Holt L, Braddock S, Troxell R, Kaplan L, Volkmar F, Klin A, Tsatsanis K, Harris DJ, Noens I, Pauls DL, Daly MJ, MacDonald ME, Morton CC, Quade BJ, Gusella JF (2008) Disruption of neurexin 1 associated with autism spectrum disorder. Am J Hum Genet 82:199-207.  Knopik VS, Smith SD, Cardon L, Pennington B, Gayan J, Olson RK, DeFries JC (2002) Differential genetic etiology of reading component processes as a function of IQ. Behav Genet 32:181-98.  Kong A and Cox NJ. (1997) Allele-sharing models: LOD scores and accurate linkage tests. Am J Hum Genet 61:1179-1188.  Krasnoperov VG, Beavis R, Chepurny OG, Little AR, Plotnikov AN, Petrenko AG (1996) The calcium-independent receptor of α-latrotoxin is not a neurexin. Biochem Biophys Res Commun 227: 868–875.  Krasnoperov VG, Bittner MA, Beavis R, Kuang Y, Salnikow KV, Chepurny OG, Little AR, Plotnikov AN, Wu D, Holz RW, Petrenko AG (1997) α-Latrotoxin stimulates exocytosis by the interaction with a neuronal G-protein-coupled receptor. Neuron 18: 925–937.  Kreienkamp HJ, Zitzer H, Gundelfinger ED, Richter D, Bockers TM (2000) The Calcium- independent Receptor for α-Latrotoxin from Human and Rodent Brains Interacts with Members of the ProSAP/SSTRIP/Shank Family of Multidomain Proteins.  J Biol Chem 275: 32387-32390.  Kriegstein AR, Noctor SC (2004) Patterns of neuronal migration in the embryonic cortex. Trends Neurosci 27:392-399.  Kruglyak L, Daly MJ, Reeve-Daly MP, Lander ES (1996) Parametric and nonparametric linkage analysis: A unified multipoint approach.  Am J Hum Genet 58:1347-1363.   106  Kubova Z, Kuba M, Peregrin J, Novakova V (1996) Visual evoked potential evidence for magnocellular system deficit in dyslexia. Physiol Res 45:87-9.  Kumar RA, Karamohamed S, Sudi J, Conrad DF, Brune C, Badner JA, Gilliam TC, Nowak NJ, Cook EH Jr, Dobyns WB, Christian SL (2008) Recurrent 16p11.2 microdeletions in autism. Hum Mol Genet 17:628-638.  LaHoste GJ, Swanson JM, Wigal SB, Glabe C, Wigal T, King N, Kennedy JL (1996) Dopamine D4 receptor gene polymorphism is associated with attention deficit hyperactivity disorder. Mol Psychiatry 1:121-124.  Lambert D, Middle F, Hamshere ML, Segurado R, Raybould R, Corvin A, Green E, O'Mahony E, Nikolov I, Mulcahy T, Haque S, Bort S, Bennett P, Norton N, Owen MJ, Kirov G, Lendon C, Jones L, Jones I, Holmans P, Gill M, Craddock N (2005) Stage 2 of the Wellcome Trust UK- Irish bipolar affective disorder sibling-pair genome screen: Evidence for linkage on chromosomes 6q16-q21, 4q12-q21, 9p21, 10p14-p12 and 18q22. Mol Psychiatry 10:831-841.  Lander E, Kruglyak L (1995) Genetic dissection of complex traits: Guidelines for interpreting and reporting linkage results. Nat Genet 11:241-247.  Lander ES, Schork NJ (1994) Genetic dissection of complex traits. Science 265:2037-2048.  Langefeld CD, Fingerlin TE (2007) Association Methods in Human Genetics. In Methods in Molecular Biology, vol 404 In Topics in Biostatistics. WT Ambrosuis, ed. pp 431-459. Totowa, NJ: Humana Press Inc.  Lawrence R, Yabe T, Hajmohammadi S, Rhodes J, McNeely M,Liu J, Lamperti ED, Toselli PA, Lech M, Spear PG, Rosenberg RD, Shworak NW (2007) The principal neuronal gD-type 3-O- sulfotransferases and their products in central and peripheral nervous system tissues. Matrix Biol 26:442-455. Leal SM, Ott J (2000) Effects of stratification in the analysis of affected sib-pair data: benefits and costs. Am J Hum Genet 66:567-575.  Lee JA, Lupski JR (2006) Genomic rearrangements and gene copy-number alterations as a cause of nervous system disorders. Neuron 52:103-121.  Lehmkuhle S, Garzia RP, Turner L, Hash T, Baro JA (1993) A defective visual pathway in children with reading disability. N Engl J Med 328:989-996.  Lelianova VG, Davletov BA, Sterling A, Rahman MA, Grishin EV, Totty NF, Ushkaryov YA (1997) α-Latrotoxin receptor, latrophilin, is a novel member of the secretin family of G protein- coupled receptors. J Biol Chem 272:21504–21508.  Lewitter FI, DeFries JC, Elston RC (1980) Genetic models of reading disability. Behav Genet 10:9- 30.  107   Liederman J, Kantrowitz L, Flannery K (2005) Male vulnerability to reading disability is not likely to be a myth: A call for new data. J Learn Disabil 38:109-129.  Light JG, Pennington BF, Gilger JW, DeFries JC (1995) Reading disability and hyperactivity: evidence for a common genetic etiology. Dev Neuropsychol 11:323–335.  Lin S, Chakravarti A, Cutler DJ (2004) Exhaustive allelic transmission disequilibrium tests as a new approach to genome-wide association studies. Nat Genet 36:1181-1188.  Lindgren SD, De Renzi E, Richman LC (1985) Cross-national comparisons of developmental dyslexia in Italy and the United States. Chil Dev 56:1404-1417.  Lintas C, Sacco R, Garbett K, Mirnics K, Militerni R, Bravaccio C, Curatolo P, Manzi B, Schneider C, Melmed R, Elia M, Pascucci T, Puglisi-Allegra S, Reichelt KL, Persico AM (2008) Involvement of the PRKCB1 gene in autistic disorder: significant genetic association and reduced neocortical gene expression. Mol Psychiatry. Mar 4 [Epub ahead of print].  Liu J, Nyholt DR, Magnussen P, Parano E, Pavone P, Geschwind D et al (2001) A genomewide screen for autism susceptibility loci. Am J Hum Genet 69:327-340.  Livingstone MS, Rosen GD, Drislane FW, Galaburda AM (1991) Physiological and anatomical evidence for a magnocellular defect in developmental dyslexia. Proc Nat Acad Sci (USA) 88:7943- 7947.  Loo SK, Fisher SE, Francks C, Ogdie MN, MacPhie IL, Yang M, McCracken JT, McGough JJ, Nelson SF, Monaco AP, Smalley SL (2004) Genome-wide scan of reading ability in affected sibling pairs with attention-deficit/hyperactivity disorder: unique and shared genetic effects. Mol Psychiatry 9:485-93.  Lovegrove W, Bowling A, Badcock D, Blackwood M (1980) Specific reading disability: differences in contrast sensitivity as a function of spatial frequency. Science 210:439-440.  Lovegrove W, Martin F, Slaghius W (1986) A theoretical and experimental case for a visual deficit in specific reading disability. Cognitive Neuropsychology 3:225-267.  Majewski H, Iannazzo L (1998) Protein kinase C: a physiological mediator of enhanced transmitter output. Prog Neurogiol 55:463-475.  Manis ER, Mcbride-Chang C, Seidenberg MS, Keating P, Doi LM, Munson B et al (1997) Are speech perception deficits associated with developmental dyslexia? J Exp Child Psychol 66:211- 235.   108  Marino C, Giorda R, Vanzin L, Nobile M, Lorusso ML, Baschirotto C, Riva L, Molteni M, Battaglia M (2004) A locus on 15q15-15qter influences dyslexia: Further support from a transmission/disequilibrium study in an italian speaking population. J Med Genet 41:42-46.  Markwardt FC (1989) Examiner’s Manual: Peabody Individual Achievement Test-Revised. American Guidelines Service: Circle Pines, MN.  Martin J, Han C, Gordon LA, Terry A, Prabhakar S, She X, Xie G, Hellsten U, Chan YM, Altherr M, Couronne O, Aerts A, Bajorek E, Black S, Blumer H, Branscomb E, Brown NC, Bruno WJ, Buckingham JM, Callen DF, Campbell CS, Campbell ML, Campbell EW, Caoile C, Challacombe JF, Chasteen LA, Chertkov O, Chi HC, Christensen M, Clark LM, Cohn JD, Denys M, Detter JC, Dickson M, Dimitrijevic-Bussod M, Escobar J, Fawcett JJ, Flowers D, Fotopulos D, Glavina T, Gomez M, Gonzales E, Goodstein D, Goodwin LA, Grady DL, Grigoriev I, Groza M, Hammon N, Hawkins T, Haydu L, Hildebrand CE, Huang W, Israni S, Jett J, Jewett PB, Kadner K, Kimball H, Kobayashi A, Krawczyk MC, Leyba T, Longmire JL, Lopez F, Lou Y, Lowry S, Ludeman T, Manohar CF, Mark GA, McMurray KL, Meincke LJ, Morgan J, Moyzis RK, Mundt MO, Munk AC, Nandkeshwar RD, Pitluck S, Pollard M, Predki P, Parson-Quintana B, Ramirez L, Rash S, Retterer J, Ricke DO, Robinson DL, Rodriguez A, Salamov A, Saunders EH, Scott D, Shough T, Stallings RL, Stalvey M, Sutherland RD, Tapia R, Tesmer JG, Thayer N, Thompson LS, Tice H, Torney DC, Tran-Gyamfi M, Tsai M, Ulanovsky LE, Ustaszewska A, Vo N, White PS, Williams AL, Wills PL, Wu JR, Wu K, Yang J, Dejong P, Bruce D, Doggett NA, Deaven L, Schmutz J, Grimwood J, Richardson P, Rokhsar DS, Eichler EE, Gilna P, Lucas SM, Myers RM, Rubin EM, Pennacchio LA (2004) The sequence and analysis of duplication- rich human chromosome 16. Nature 432:988-994.  Matsushita H, Lelianova VG, Ushkaryov YA (1999) The latrophilin family: multiply spliced G protein-coupled receptors with differential tissue distribution. FEBS Lett  443: 348–352.  McAnally KI, Stein JF (1996) Auditory temporal coding in dyslexia. Proc Roy Acad Lond-Series B: Biol Sci 263:961-965.  McGee R, Prior M, Willams S, Smart D, Sanson A (2002) The long-term significance of teacher- rated hyperactivity and reading ability in childhood: findings from two longitudinal studies. J Child Psychol Psychiatry 43:1004–1017.  Meng H, Smith SD, Hager K, Held M, Liu J, Olson RK, Pennington BF, DeFries JC, Gelernter J, O'Reilly-Pol T, Somlo S, Skudlarski P, Shaywitz SE, Shaywitz BA, Marchione K, Wang Y, Paramasivam M, LoTurco JJ, Page GP, Gruen JR (2005) DCDC2 is associated with reading disability and modulates neuronal development in the brain. Proc Natl Acad Sci USA 102:17053-17058.  Miller SA, Dykes DD, Polesky HF (1988) A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res 16:1215.   109  Miyoshi K, Asanuma M, Miyazaki I, Diaz-Corrales FJ, Katayama T, Tohyama M, Ogawa N (2004) DISC1 localizes to the centrosome by binding to kendrin. Biochem Biophys Res Commun 317:1195-1199.  Morgan D, Eley L, Sayer J, Strachan T, Yates LM, Craighead AS, Goodship JA (2002) Expression analyses and interaction with the anaphase promoting complex protein Apc2 suggest a role of inversin in primary cilia and involvement in the cell cycle. Hum Mol Genet 11: 3345- 3350.  Morris DW, Robinson L, Turic D, Duke M, Webb V, Southgate L, Hamshere M, Williams N, McGuffin P, Rowen L, Stevenson J, Krawczak M, Owen MJ, Williams J, O’Donovan MC (2000) Linkage disequilibrium and positional candidate gene analyses of a susceptibility locus for reading disability on chromosome 15q. Am J Med Genet 96:471.  Morris NR, Efimov VP, Xiang X (1998) Nuclear migration, nucleokinesis and lissencephaly. Treds Cell Biol 8:467-470.  Muglia P, Jain U, Macciardi F, Kennedy JL (2000) Adult attention deficit hyperactivity disorder and the dopamine D4 receptor gene. Am J Med Genet 96:273-277.  Nicolson RI, Fawcett AJ, Dean P (2001) Developmental dyslexia: The cerebellar deficit hypothesis. Trends Neurosci 24:508-511.  Nopola-Hemmi J, Taipale M, Haltia T, Lehesjoki AE, Voutilainen A, Kere J (2000) Two translocations of chromosome 15q associated with dyslexia. J Med Genet  37:771-775.  Nopola-Hemmi J, Myllyluoma B, Haltia T, Taipale M, Ollikainen V, Ahonen T, Voutilainen A, Kere J, Widen E (2001) A dominant gene for developmental dyslexia on chromosome 3. J Med Genet 38:658-664.  Nothen MM, Schulte-Korne G, Grimm T, Cichon S, Vogt IR, Muller-Myhsok B, Propping P, Remschmidt H (1999) Genetic linkage analysis with dyslexia: evidence for linkage of spelling disability to chromosome 15. Eur Child Adolesc Psychiatry 8 (Suppl 3):56-9.  Ogdie MN, Macphie IL, Minassian SL, Yang M, Fisher SE, Francks C, Cantor RM, McCracken JT, McGough JJ, Nelson SF, Monaco AP, Smalley SL (2003) A genomewide scan for attention- deficit/hyperactivity disorder in an extended sample: suggestive linkage on 17p11. Am J Hum Genet 72:1268-79.  Ogdie MN, Fisher SE, Yang M, Ishii J, Francks C, Loo SK, Cantor RM, McCracken JT, McGough JJ, Smalley SL, Nelson SF (2004) Attention deficit hyperactivity disorder: fine mapping supports linkage to 5p13, 6q12, 16p13, and 17p11. Am J Hum Genet 75:661-8.  Olson R, Wise B, Conners F, Rack J, Fulker D (1989) Specific deficits in component reading and language skills: genetic and environmental influences. J Learn Dis 22:339-348.  110   Paracchini S, Thomas A, Castro S, Lai C, Paramasivam M, Wang Y, Keating BJ, Taylor JM, Hacking DF, Scerri T, Francks C, Richardson AJ, Wade-Martins R, Stein JF, Knight JC, Copp AJ, Loturco J, Monaco AP (2006) The chromosome 6p22 haplotype associated with dyslexia reduces the expression of KIAA0319, a novel gene involved in neuronal migration. Hum Mol Genet 15:1659-1666.  Paul I, Bott C, Heim S, Eulitz C, Elbert T (2006a) Reduced hemispheric asymmetry of the auditory N260m in dyslexia. Neuropsychologia 44:785-94.  Paul I, Bott C, Heim S, Wienbruch C, Elbert TR (2006b) Phonological but not auditory discrimination is impaired in dyslexia. Eur J Neurosci 24:2945-2953.  Paulesu E, Demonet JF, Fazio F, McCrory E, Chanoine V, Brunswick N, Cappa SF, Cossu G, Habib M, Frith CD, Frith U (2001) Dyslexia: cultural diversity and biological unity. Science 291:2165-2167.  Pearson CG, Culver BP, Winey M (2007) Centrioles Want to Move Out and Make Cilia. Dev Cell 13:319-321.  Pennington BF, Lefly DL (2001) Early reading development in children at family risk for dyslexia. Child Dev 72:816-833.  Pennington BF, Gilger JW, Pauls D, Smith SA, Smith SD, DeFries JC (1991) Evidence for major gene transmission of developmental dyslexia. JAMA 266:1527-34.  Petryshen T, Kaplan B, Liu MF, Field L (2000) Absence of significant linkage between phonological coding dyslexia and chromosome 6p23-p21.3, as determined by use of quantitative-trait methods: confirmation of qualitative analyses. Amer J Hum Genet 66:708-714.  Petryshen TL, Kaplan BJ, Liu MF, Schmill de French N, Tobias R, Hughes M, Field LL (2001) A susceptibility locus (DYX4) on chromosome 6q influencing phonological coding dyslexia. Neuropsychiatric Genet 105:507-517.  Petryshen TL, Kaplan BJ, Hughes ML, Tzenova J, Field LL (2002) Supportive evidence for the DYX3 dyslexia gene in Canadian families. J Med Genet (Neuropsychiatric Genet) 39:125-126.  Peyrard-Janvid M, Anthoni H, Onkamo P, Lahermo P, Zucchelli M, Kaminen N, Hannula- Jouppi K, Nopola-Hemmi J, Voutilainen A, Lyytinen H, Kere J (2004) Fine mapping of the 2p11 dyslexia locus and exclusion of TACR1 as a candidate gene. Hum Genet 114:510-516.  Philippi A, Roschmann E, Tores F, Lindenbaum P, Benajou A, Germain-Leclerc L, Marcaillou C, Fontaine K, Vanpeene M, Roy S, Maillard S, Decaulne V, Saraiva JP, Brooks P, Rousseau F, Hager J (2005) Haplotypes in the gene encoding protein kinase c-beta (PRKCB1) on chromosome 16 are associated with autism. Mol Psychiatry 10:950-960.  111   Pilz D, Stoodley N, Golden JA (2002) Neuronal migration, cerebral cortical development, and cerebral cortical anomalies. J Neuropathol Exp Neurol 61:1-11.  Ploughman L, Boehnke M (1989) Estimating the power of a proposed linkage study for a complex genetic trait. Am J Hum Genet 44:543-551.  Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira, MAR, Bender D, Maller, J, Sklar P, de Bakker PIW, Daly MJ, Sham PC (2007) PLINK: A Tool Set for Whole-Genome Association and Population-Based Linkage Analyses. Am Hum Genet 81:559 – 575.  Purvis KL, Tannock R (1997) Language abilities in children with attention deficit hyperactivity disorder, reading disabilities, and normal controls. J Abnorm Child Psychol 25:133-144.  Rabin M, Wen XL, Hepburn M, Lubs HA, Feldman E, Duara R (1993) Suggestive linkage of developmental dyslexia to chromosome 1p34-p36. Lancet 342:178.  Ramus F (2004) Neurobiology of dyslexia: A reinterpretation of the data. Trends Neurosci 27:720-726.  Raskind WH, Igo RP, Chapman NH, Berninger VW, Thomson JB, Matsushita M, Brkanac Z, Holzman T, Brown M, Wijsman EM (2005)  A genome scan in multigenerational families with dyslexia: Identification of a novel locus on chromosome 2q that contributes to phonological decoding efficiency. Mol Psychiatry 10:699-711.  Risch N, Merikangas K (1996) The future of genetic studies of complex human diseases. Science 273:1516-1517.  Roberts SB, MacLean CJ, Neale MC, Eaves LJ, Kendler KS (1999) Replication of linkage studies of complex traits: an examination of variation in location estimates. Am J Hum Genet 65:876-884.  Rochell K, Talcott J (2006) Impaired balance in developmental dyslexia? A meta-analysis of the contending evidence. J Child Psychol Psychiatry 47: 1159–1166.  Rosner J, Simon DP (1971) The auditory analysis test: an initial report. J Learn Dis 4:384-392.  Rowe DC, Stever C, Giedinghagen LN, Gard JM, Cleveland HH, Terris ST, Mohr JH, Sherman S, Abramowitz A, Waldman ID (1998) Dopamine DRD4 receptor polymorphism and attention deficit hyperactivity disorder. Mol Psychiatry 3:419-426.  Rumsey JM (1992) The biology of developmental dyslexia. JAMA 19:912-915   112  Rumsey JM, Donohue BC, Brady DR, Nace K, Giedd JN, Andreason P (1997) A magnetic resonance imaging study of planum temporale asymmetry in men with developmental dyslexia. Arch Neurol  54:1481-1489.  Rutter M (1978) Prevalence and types of dyslexia. In: Benton A, Pearl D (eds) Dyslexia: an Appraisal of Current Knowledge. Oxford University Press, New York, NY, pp 5-28.  Rutter M, Caspi A, Fergusson D, Horwood LJ, Goodman R, Maughan B, Moffitt TE, Meltzer H, Carroll J (2004) Sex differences in developmental reading disability: New findings from 4 epidemiological studies. JAMA 291:2007-2012.  Salmelin R, Service E, Kiesila P, Uutela K, Salonen O (1996) Impaired visual word processing in dyslexia revealed with magnetoencephalography. Ann Neurol 40:157-162.  Schulte-Körne G, Bartling J, Deimel W, et al (2004) Motion-onset VEPs in dyslexia. Evidence for visual perceptual deficit. Neuroreport 15:1075–1078  Schulte-Körne G, Deimel W, Bartling J, Remschmidt H (1999) Pre-attentive processing of auditory patterns in dyslexic human subjects. Neruosci Lett 276: 41-44.  Schulte-Körne G, Grimm T, Nöthen MM, Müller-Myhsok B, Cichon S, Vogt IR, Propping P, Remschmidt H (1998) Evidence for linkage of spelling disability to chromosome 15. Am J Hum Genet 63:279-82.  Schumacher J, Anthoni H, Dahdouh F, Konig IR, Hillmer AM, Kluck N, Manthey M, Plume E, Warnke A, Remschmidt H, Hulsmann J, Cichon S, Lindgren CM, Propping P, Zucchelli M, Ziegler A, Peyrard-Janvid M, Schulte-Korne G, Nothen MM, Kere J (2006) Strong genetic evidence of DCDC2 as a susceptibility gene for dyslexia. Am J Hum Genet 78:52-62.  Sebat J, Lakshmi B, Malhotra D, Troge J, Lese-Martin C, Walsh T, Yamrom B, Yoon S, Krasnitz A, Kendall J, Leotta A, Pai D, Zhang R, Lee YH, Hicks J, Spence SJ, Lee AT, Puura K, Lehtimäki T, Ledbetter D, Gregersen PK, Bregman J, Sutcliffe JS, Jobanputra V, Chung W, Warburton D, King MC, Skuse D, Geschwind DH, Gilliam TC, Ye K, Wigler M (2007) Strong association of de novo copy number mutations with autism. Science 316:445-449.  Sebat J, Lakshmi B, Troge J, Alexander J, Young J, Lundin P, Månér S, Massa H, Walker M, Chi M, Navin N, Lucito R, Healy J, Hicks J, Ye K, Reiner A, Gilliam TC, Trask B, Patterson N, Zetterberg A, Wigler M (2004) Large-scale copy number polymorphism in the human genome. Science 305:525-528.  Seeger M, Tear G, Ferres-Marco D, Goodman CS (1993) Mutations affecting growth cone guidance in drosophila: Genes necessary for guidance toward or away from the midline. Neuron 10:409-426.   113  Segalowitz, S.J., B.E. Berge, S. Lawson and D. Brown. 1994. If you can replicate the handednessimmune disorder effect, the more power to you. Brain Cogn. 26:217-227.  Seshadri S, DeStefano AL, Au R, Massaro JM, Beiser AS, Kelly-Hayes M, Kase CS, D'Agostino RB Sr, Decarli C, Atwood LD, Wolf PA (2007) Genetic correlates of brain aging on MRI and cognitive test measures: a genome-wide association and linkage analysis in the Framingham Study. BMC Med Genet 8:S1-S15.  Shapleske J, Rosselll SL, Woodruff PWR, David AS (1999) The planum temporale: a systematic, quantitative review of its structure, functional and clinical significance. Brain Res Rev 29:26-49.  Shaywitz BA, Shaywitz SE, Pugh KR, Constable RT, Skudlarski P, Fulbright RK, Bronen RA, Fletcher JM, Shankweiler DP, Katz L, et al (1995) Sex differences in the functional organization of the brain for language. Nature  373:607-609.  Shaywitz SE, Shaywitz BA, Fletcher JM, Escobar MD (1990) Prevalence of reading disability in boys and girls: results of the Connecticut Longitudinal Study. JAMA 264:998-1002.  Shaywitz SE, Escobar MD, Shaywitz BA, Fletcher JM, Makuch R (1992) Evidence that dyslexia may represent the lower tail of a normal distribution of reading ability. N Engl J Med 326:145- 150.  Shaywitz SE (1996) Dyslexia. Scientific American, pp 98-104.  Shaywitz SE, Shaywitz BA, Pugh KR, Fulbright RK, Constable RT, Mencl WE, Shankweiler DP, Liberman AM, Skudlarski P, Fletcher JM, Katz L, Marchione KE, Lacadie C, Gatenby C, Gore JC (1998) Functional disruption in the organization of the brain for reading in dyslexia. Proc Natl Acad Sci USA 95:2636-2641.  Shaywitz SE (1998) Dyslexia. N Engl J Med 338:307-312.  Shaywitz SE, Shaywitz BA, (2003) The science of reading and dyslexia. J AAPOS 7:158-166.  Smalley SL, Bailey JN, Palmer CG, Cantwell DP, McGough JJ, Del'Homme MA, Asarnow JR, Woodward JA, Ramsey C, Nelson SF (1998) Evidence that the dopamine D4 receptor is a susceptibility gene in attention deficit hyperactivity disorder. Mol Psychiatry 3:427-430.  Smalley SL, Kustanovich V, Minassian SL, Stone JL, Ogdie MN, McGough JJ, McCracken JT, MacPhie IL, Francks C, Fisher SE, Cantor RM, Monaco AP, Nelson SF (2002) Genetic linkage of attention-deficit/hyperactivity disorder on chromosome 16p13, in a region implicated in autism. Am J Hum Genet 71:959-963.   114  Smalley SL, Loo SK, Yang MH, Cantor RM (2005)  Toward localizing genes underlying cerebral asymmetry and mental health. Am J Med Genet B Neuropsychiatr Genet135:79-84. Erratum in: Am J Med Genet B Neuropsychiatr Genet136:107.  Smith SD, Kimberling WJ, Pennington BF (1991) Screening for multiple genes influencing dyslexia. Reading Writing Interdiscipl J 3:285-298.  Smith SD, Kimberling WJ, Pennington BF, Lubs HA (1983) Specific reading disability: Identification of an inherited from through linkage anlysis. Science 219:1345-1347.  Snowling MJ (1991) Developmental reading disorders. J Child Psychol Psychiatry 32:49-77.  Snowling M (2000) Dyslexia. Blackwell, Oxford.  Spektor A, Tsang WY, Khoo D, Dynlacht BD (2007) Cep97 and CP110 suppress a cilia assembly program. Cell 130:678-90.  Spielman RS, McGinnis RE, Ewens WJ (1993) Transmission test for linkage disequilibrium: The insulin gene region and insulin-dependent diabetes mellitus (IDDM).  Am J Hum Genet 52:506-516.  Spielman RS, Ewens WJ (1996) The TDT and other family-based tests for linkage disequilibrium and association. Am J Hum Genet 59:983-989.  Stanovich KE (1988) Explaining the differences between the dyslexic and the garden-variety poor reader: the phonological-core variable-difference model. J Learn Dis 21:590-604.  Stein J (2001) The magnocellular theory of developmental dyslexia. Dyslexia 7:12-36.  Stein J, Walsh V (1997) To see but not to read; the magnocellular theory of dyslexia. Trends Neurosci 20:147-152.  Stein JF (1993) Visuospatial perception in disabled readers. In: Willows DM, Kruk RS, Corcos E (eds).Visual Prosesses in Reading and Reading Disabilities. Hove, London, 331-346.  Stevenson J, Pennington BF, Gilger JW, DeFries JC, Gillis JJ (1993) Hyperactivity and spelling disability: testing for shared genetic aetiology. J Child Psychol Psychiatry 34:1137–1152.  Sugita S, Ichtchenko K, Khvotchev M, Südhof TC (1998) alpha-Latrotoxin receptor CIRL/latrophilin 1 (CL1) defines an unusual family of ubiquitous G-protein-linked receptors. G- protein coupling not required for triggering exocytosis. J Biol Chem 273: 32715-32724.  Sunohara GA, Roberts W, Malone M, Schachar RJ, Tannock R, Basile VS, Wigal T, Wigal SB, Schuck S, Moriarty J, Swanson JM, Kennedy JL, Barr CL (2000) Linkage of the dopamine D4  115  receptor gene and attention-deficit/hyperactivity disorder. J Am Acad Child Adolesc Psychiatry 39:1537-1542.  Swanson JM, Sunohara GA, Kennedy JL, Regino R, Fineberg E, Wigal T, Lerner M, Williams L, LaHoste GJ, Wigal S (1998) Association of the dopamine receptor D4 (DRD4) gene with a refined phenotype of attention deficit hyperactivity disorder (ADHD): a family-based approach. Mol Psychiatry 3:38-41.  Szatmari P, Paterson AD, Zwaigenbaum L, Roberts W, Brian J, Liu XQ, Vincent JB, Skaug JL, Thompson AP, Senman L, et al (2007) Mapping autism risk loci using genetic linkage and chromosomal rearrangements. Nat Genet 39:319 328.  Tallal P (1980) Auditory temporal perception, phonics, and reading disabilities in children. Brain and Language 9:182-198.  Tallal P, Piercy M (1975) Developmental aphasia: the perception of brief vowels and extended stop consonants. Neuropsychologia 13:69-74.  Tallal P, Stark RE (1982) Perceptual/motor profiles of reading impaired children with or without concomitant oral language deficits. Ann Dyslexia 32:163-176.  Tallal P, Miller S, Fitch RH (1993) Neurobiological basis of speech: a case for the preeminance of temporal processing. Ann N Y Acad Sci 682:27-47.  Taipale M, Kaminen N, Nopola-Hemmi J, Haltia T, Myllyluoma B, Lyytinen H, Muller K, Kaaranen M, Lindsberg PJ, Hannula-Jouppi K, Kere J (2003) A candidate gene for developmental dyslexia encodes a nuclear tetratricopeptide repeat domain protein dynamically regulated in brain. Proc Natl Acad Sci U S A 100:11553-11558.  Tanaka T, Serneo FF, Higgins C, Gambello MJ, Wynshaw-Boris A, Gleeson JG (2004) Lis1 and doublecortin function with dynein to mediate coupling of the nucleaus to the centrosome in neuronal migration. J Cell Biol 165:709-721.  Temple E, Deutsch GK, Poldrack RA, Miller SL,Tallal P, Merzenich MM, Gabrieli JD (2003) Neural deficits in children with dyslexia ameliorated by behavioral remediation: Evidence from functional MRI. Proc Natl Acad Sci USA 100:2860-2865.  Thomson G (1995) Mapping disease genes: Family-based association studies.  Am J Hum Genet 57:487-498.  Toivonen HT, Onkamo P, Vasko K, Ollikainene V, Sevon P, Mannila H, Herr M, Kere J (2000) Data mining applied to linkage disequilibrium mapping. Am J Hum Genet 67:133-145.  Turic D, Langley K, Mills S, Stephens M, Lawson D, Govan C, Williams N, Van Den Bree M, Craddock N, Kent L, Owen M, O'Donovan M, Thapar A (2004) Follow-up of genetic linkage  116  findings on chromosome 16p13: evidence of association of N-methyl-D aspartate glutamate receptor 2A gene polymorphism with ADHD. Mol Psychiatry 9:169-73.  Turic D, Morris DW, Robinson L, Duke M, Grierson A, Raha-Chowdhury R, Southgate L, Webb V, Hamshere M, Williams B, McGuffin P, Stevenson J, Krawczak M, Owen MJ, O’Donovan MC, Williams J (2000) Linkage disequilibrium mapping provides evidence for a susceptibility gene for reading disability on chromosome 6p. Am J Med Genet 96:556.  Turic D, Robinson L, Duke M, Morris DW, Webb V, Hamshere M, Milham C, Hopkin E, Pound K, Fernando S, Grierson A, Easton M, Williams N, Van Den Bree M, Chowdhury R, Gruen J, Stevenson J, Krawczak M, Owen MJ, O'Donovan MC, Williams J (2003) Linkage disequilibrium mapping provides further evidence of a gene for reading disability on chromosome 6p21.3-22. Mol Psychiatry 8:176-85.  Tzenova J, Kaplan BJ, Petryshen TL, Field LL (2004) Confirmation of a dyslexia susceptibility locus on chromosome 1p34-p36 in a set of 100 Canadian families. Am J Med Genet B Neuropsychiatr Genet 127:117-124.  Ueda M, Graf R, MacWilliams HK, Schliwa M, Euteneuer U (1997) Centrosome positioning and directionality of cell movements. Proc Natl Acad Sci USA 94:9674-9678.  Ulgen A, Li W (2005) Comparing single-nucleotide polymorphism marker-based and microsatellite marker-based linkage analyses. BMC Genetics 6(Suppl 1):S13.  Ushkaryov YA, Petrenko AG, Geppert M, Sudhof TC (1992) Neurexins: synaptic cell surface proteins related to the α-latrotoxin receptor and laminin. Science 257:50–56.  Ushkaryov YA, Volynski KE, Ashton AC. (2004) The multiple actions of black widow spider toxins and their selective use in neurosecretion studies. Toxicon. 43:527-542.  Van Orden GC, Goldinger SD (1996) Phonologic mediation in skeilled and dyslexic reading. In: Chase CH, Rosen GD, Sherman GF (eds). Developmental dyslexia: neural, cognitive, and genetic mechanisms. York Press, Baltimore, MD, pp 185-223.  Vogel SA (1990) Gender differences in intelligence, language, visual-motor abilities and academic achievement in males and females with learning disabilities: a review of the literature. J Learning Disabilities 23:44-52. Walsh T, McClellan JM, McCarthy SE, Addington AM, Pierce SB, Cooper GM, Nord AS, Kusenda M, Malhotra D, Bhandari A, Stray SM, Rippey CF, Roccanova P, Makarov V, Lakshmi B, Findling RL, Sikich L, Stromberg T, Merriman B, Gogtay N, Butler P, Eckstrand K, Noory L, Gochman P, Long R, Chen Z, Davis S, Baker C, Eichler EE, Meltzer PS, Nelson SF, Singleton AB, Lee MK, Rapoport JL, King MC, Sebat J. (2008) Rare Structural Variants Disrupt Multiple Genes in Neurodevelopmental Pathways in Schizophrenia.  Science [March 27, 2008 Epub ahead of print].   117  Wechsler D (1974) Manual for the Wechsler intelligence scale for children - revised. Psychological Corporation, New York.  Wechsler D (1981) Wechsler adult intelligence scale - revised. Psychological Corporation, New York.  Weiss LA, Shen Y, Korn JM, Arking DE, Miller DT, Fossdal R, Saemundsen E, Stefansson H, Ferreira MA, Green T, Platt OS, Ruderfer DM, Walsh CA, Altshuler D, Chakravarti A, Tanzi RE, Stefansson K, Santangelo SL, Gusella JF, Sklar P, Wu BL, Daly MJ; Autism Consortium (2008) Association between microdeletion and microduplication at 16p11.2 and autism. N Engl J Med 358:667-675. Wellcome Trust Case Control Consortium (2007) Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature 447:661-678.  Wigg KG, Couto JM, Feng Y, Anderson B, Cate-Carter TD, Macciardi F, Tannock R, Lovett MW, Humphries TW, Barr CL (2004) Support for EKN1 as the susceptibility locus for dyslexia on 15q21. Mol Psychiatry 9:1111-1121.  Wigginton JE, Cutler DJ, Abecasis GR (2005) A Note on Exact Tests of Hardy-Weinberg Equilibrium. Am J Hum Genet 76: 887-893.  Wigginton JE, Abecasis GR (2005) PEDSTATS: descriptive statistics, graphics and quality assessment for gene mapping data. Bioinformatics 21:3445-3447.  Willcutt EG, Pennington BF, DeFries JC (2000) Twin study of the etiology of comorbidity between reading disability and attention-deficit/hyperactivity disorder. Am J Med Genet 96:293- 301.  Willcutt EG, Pennington BF, Boada R, Ogline JS, Tunick RA, Chhabildas NA, Olson RK (2001) A comparison of the cognitive deficits in reading disability and attention-deficit/hyperactivity disorder. J Abnorm Psychol 110:157-172.  Willcutt EG, Pennington BF, Olson RK, Chhabildas N, Hulslander J (2005)  Neuropsychological analyses of comorbidity between reading disability and attention deficit hyperactivity disorder: in search of the common deficit. Dev Neuropsychol 27:35-78. Willcutt EG, Pennington BF, Olson RK, DeFries JC (2007) Understanding comorbidity: a twin study of reading disability and attention-deficit/hyperactivity disorder. Am J Med Genet B Neuropsychiatr Genet. 144:709-714.  Williams J, O'Donovan MC (2006) The genetics of developmental dyslexia. Eur J Hum Genet 14:681-689.  Wiltshire S, Bell JT, Groves CJ, Dina C, Hattersley AT, Frayling TM, Walker M, Hitman GA, Vaxillaire M, Farrall M, Froguel P, McCarthy MI (2006) Epistasis between type 2 diabetes  118  susceptibility Loci on chromosomes 1q21-25 and 10q23-26 in northern Europeans. Ann Hum Genet 70:726-37.  Witton C, Talcott JB, Hansen PC, Richardson AJ, Griffiths TD, Rees A, Stein JF, Green GGR (1998) Sensitivity to dynamic auditory and visual stimuli predicts nonword reading ability in both dyslexic and normal readers. Curr Biol 8:791-797.  Wingo AP, Ghaemi SN (2007) A systematic review of rates and diagnostic validity of comorbid adult attention-deficit/hyperactivity disorder and bipolar disorder. J Clin Psychiatry 68:1776- 1784.  Wolf M and Bowers PG (1999) The double-deficit hypothesis for the developmental dyslexias. J Educ Psychol 91:415-438.  Wolff PH, Melngailis I (1994) Family patterns of developmental dyslexia: clinical findings. Amer J Med Genet (Neuropsychiatric Genetics) 54:122-131.  Woodcock RW (1987) Woodcock Reading Mastery Tests. American Guidance Service, Circle Pines, MN.  Woodcock RW, Johnson MB (1989) Woodcock-Johnson Psycholoeducational Battery -Revised. DLM Teaching Resources, Allen, Texas.  Wu J, Song TB, Li YJ, He KS, Ge L, Wang LR (2007) Prenatal restraint stress impairs learning and memory and hippocampal PKCbeta1 expression and translocation in offspring rats. Brain Res 1141: 205-213.  Wu W, Wong K, Chen J, Jiang Z, Dupuis S, Wu JY, Rao Y (1999) Directional guidance of neuronal migration in the olfactory system by the protein slit. Nature 400:331-336.  Yokoyama T, Copeland NG, Jenkins NA, Montgomery CA, Elder FF, Overbeek PA. (1993) Reversal of left-right asymmetry: a situs inversus mutation. Science 260:679-682. Yule W, Rutter M (1976) Epidemiological and social implications of specific reading retardation. In: Knights RM, Bakker KJ (eds) The Neuropsychology of Learning Disorders: Theoretical Approaches. University Park Press, Baltimore, MD.  Zhu Y, Li H, Zhou L, Wu JY, Rao Y (1999) Cellular and molecular guidance of GABAergic neuronal migration from an extracortical origin to the neocortex. Neuron 23:473 485.  119  APPENDIX1 Table 13 - 2-SNP haplotype TDT p-values for all haplotypes in a window when one of the two markers was associated with PCD in the single marker association tests for the 16p12 region. A,C,G,T alleles are represented by  1,2,3,4. P values <0.05 are in bold. T= number of transmitted haplotypes, U= number of untransmitted haplotypes  Window  Haplotype  T  U  CHISQ P  SNPs  WIN30  13  47  29 4.263 0.0390 rs2125191  rs9934960  WIN30  22  14  14 0 1.0000 rs2125191  rs9934960  WIN30  12  35  53 3.682 0.0550 rs2125191  rs9934960  WIN70  31  47.92  27.84 5.318 0.0211 rs7202362  rs4782209  WIN70  23  2.229  5.915 1.669 0.1964 rs7202362  rs4782209  WIN70  33  1.771  4.771 1.376 0.2408 rs7202362  rs4782209  WIN70  21  27.39  40.77 2.629 0.1049 rs7202362  rs4782209  WIN74  23  23  15 1.684 0.1944 rs8051503  rs1023442  WIN74  43  24  16 1.6 0.2059 rs8051503  rs1023442  WIN74  41  25  41 3.879 0.0489 rs8051503  rs1023442  WIN75  13  25.82  47.32 6.318 0.0120 rs1023442  rs7194009  WIN75  31  9.686  6.367 0.686 0.4075 rs1023442  rs7194009  WIN75  11  28.57  21.35 1.043 0.3070 rs1023442  rs7194009  WIN75  33  30.35  19.4 2.414 0.1202 rs1023442  rs7194009  WIN81  23  30  13 6.721 0.0095 rs11640077 rs12447418  WIN81  21  13  18 0.807 0.3692 rs11640077 rs12447418  WIN81  41  28  40 2.118 0.1456 rs11640077 rs12447418  WIN82  32  5  4 0.111 0.7389 rs12447418 rs12927277  WIN82  33  29  11 8.1 0.0044 rs12447418 rs12927277  WIN82  12  14  33 7.681 0.0056 rs12447418 rs12927277  WIN94  44  20  8 5.143 0.0233 rs1031634  rs2608173  WIN94  22  31  30 0.016 0.8981 rs1031634  rs2608173  WIN94  42  35  48 2.036 0.1536 rs1031634  rs2608173  WIN95  44  20  7 6.259 0.0124 rs2608173  rs889201  WIN95  22  31  30 0.016 0.8981 rs2608173  rs889201  WIN95  24  33  47 2.45 0.1175 rs2608173  rs889201  WIN139  43  4.639  5.639 0.097 0.7551 rs1862659  rs9939346  WIN139  23  13.36  24 3.029 0.0818 rs1862659  rs9939346  WIN139  21  20.28  9 4.344 0.0372 rs1862659  rs9939346  WIN140  14  2.003  0.502 0.9 0.3427 rs9939346  rs10521122  WIN140  34  10.49  27.99 7.959 0.0048 rs9939346  rs10521122  WIN140  12  17.49  7.495 3.999 0.0455 rs9939346  rs10521122  WIN140  32  31  25 0.644 0.4224 rs9939346  rs10521122   120   Table 13 - continued  Window  Haplotype  T  U  CHISQ P  SNPs  WIN141  23  22.68  30.31 1.099 0.2944 rs10521122 rs4783488  WIN141  22  50.17  28.17 6.178 0.0129 rs10521122 rs4783488  WIN141  42  13.65  26.5 4.111 0.0426 rs10521122 rs4783488  WIN171  43  24.8  11.9 4.535 0.0332 rs8051531  rs9940479  WIN171  21  24.91  41.76 4.256 0.0391 rs8051531  rs9940479  WIN171  41  27.99  25.04 0.164 0.6859 rs8051531  rs9940479  WIN172  11  34  22 2.571 0.1088 rs9940479  rs120908  WIN172  12  25  53 10.05 0.0015 rs9940479  rs120908  WIN172  31  25  10 6.429 0.0112 rs9940479  rs120908  WIN173  14  29.02  15.45 4.139 0.0419 rs120908  rs2340988  WIN173  24  23.99  39.88 3.951 0.0468 rs120908  rs2340988  WIN173  13  18.87  11.99 1.532 0.2158 rs120908  rs2340988  WIN173  23  30.12  34.68 0.32 0.5714 rs120908  rs2340988   121  APPENDIX2 Table 14 - 2-SNP haplotype TDT p-values for all haplotypes in a window when one of the two markers was associated with PCD in the single marker association tests for the 2p21 region. A,C,G,T alleles are represented by  1,2,3,4. P values <0.05 are in bold. T= number of transmitted haplotypes, U= number of untransmitted haplotypes  Window  Haplotype  T  U  CHISQ  P  SNPs  WIN46  11  26  28 0.07407 0.7855 rs6736894  rs2004578  WIN46  41  26  42 3.765 0.0524 rs6736894  rs2004578  WIN46  42  46  28 4.378 0.0364 rs6736894  rs2004578  WIN47  14  34  45 1.532 0.2159 rs2004578  rs7590844  WIN47  12  11  18 1.69 0.1936 rs2004578  rs7590844  WIN47  22  46  28 4.378 0.0364 rs2004578  rs7590844  WIN49  42  26  38 2.25 0.1336 rs10176316 rs4953699  WIN49  22  40  45 0.2941 0.5876 rs10176316 rs4953699  WIN49  23  45  28 3.959 0.0466 rs10176316 rs4953699  WIN51  23  8.442  22.04 6.063 0.0138 rs4953706  rs10182679  WIN51  43  15.72  17.28 0.07432 0.7852 rs4953706  rs10182679  WIN51  21  20.51  16.87 0.3556 0.5510 rs4953706  rs10182679  WIN51  41  41.27  29.75 1.867 0.1719 rs4953706  rs10182679  WIN52  11  27.07  25.44 0.05054 0.8221 rs10182679 rs4608578  WIN52  32  12.81  17.69 0.7816 0.3766 rs10182679 rs4608578  WIN52  31  9.992  23.85 5.678 0.0172 rs10182679 rs4608578  WIN52  12  42.48  25.37 4.318 0.0377 rs10182679 rs4608578  WIN72  11  2.585  2.571 3.83E‐05 0.9951 rs13412979 rs17335631  WIN72  22  14.54  19.54 0.7335 0.3918 rs13412979 rs17335631  WIN72  12  35.18  50.72 2.811 0.0936 rs13412979 rs17335631  WIN72  21  40.7  20.18 6.92 0.0085 rs13412979 rs17335631  WIN73  14  30  19 2.469 0.1161 rs17335631 rs6748247  WIN73  24  20  45 9.615 0.0019 rs17335631 rs6748247  WIN73  12  23  9 6.125 0.0133 rs17335631 rs6748247  WIN74  44  28  39 1.806 0.1790 rs6748247  rs12616572  WIN74  41  34  35 0.01449 0.9042 rs6748247  rs12616572  WIN74  21  22  10 4.5 0.0339 rs6748247  rs12616572  WIN88  41  31  26 0.4386 0.5078 rs11898901 rs10190161  WIN88  42  43  27 3.657 0.0558 rs11898901 rs10190161  WIN88  22  30  51 5.444 0.0196 rs11898901 rs10190161  WIN93  31  25.5  43.5 4.696 0.0302 rs7558957  rs7583940  WIN93  14  38.5  28.5 1.493 0.2218 rs7558957  rs7583940  WIN93  34  26.5  18.5 1.422 0.2330 rs7558957  rs7583940   122   Table 14 - continued  Window  Haplotype  T  U  CHISQ  P  SNPs  WIN94  13  20.67  38.4 5.322 0.0211 rs7583940  rs10202948  WIN94  43  47  27.82 4.915 0.0266 rs7583940  rs10202948  WIN94  11  9.824  10.27 0.009863 0.9209 rs7583940  rs10202948  WIN94  41  14.76  15.76 0.03276 0.8564 rs7583940  rs10202948  WIN152  24  28  54 8.244 0.0041 rs748573  rs7576649  WIN152  22  38  21 4.898 0.0269 rs748573  rs7576649  WIN152  34  32  23 1.473 0.2249 rs748573  rs7576649  WIN153  44  29.14  52.96 6.907 0.0086 rs7576649  rs11884064  WIN153  22  1.115  1.145 0.000397 0.9841 rs7576649  rs11884064  WIN153  42  33.95  23.7 1.822 0.1771 rs7576649  rs11884064  WIN153  24  33.69  20.09 3.437 0.0637 rs7576649  rs11884064  WIN159  11  12.88  19.86 1.486 0.2229 rs1522987  rs4557033  WIN159  31  27.05  39.93 2.477 0.1155 rs1522987  rs4557033  WIN159  33  46.74  27.93 4.735 0.0296 rs1522987  rs4557033  WIN162  33  11.44  20.58 2.61 0.1062 rs7577273  rs10865208  WIN162  13  37.03  16.31 8.053 0.0045 rs7577273  rs10865208  WIN162  31  21.89  23.51 0.05726 0.8109 rs7577273  rs10865208  WIN162  11  29.52  39.49 1.441 0.2300 rs7577273  rs10865208  WIN163  33  38.49  21.38 4.89 0.0270 rs10865208 rs4953273  WIN163  32  2.54  5.284 0.9625 0.3266 rs10865208 rs4953273  WIN163  12  10.46  20.87 3.459 0.0629 rs10865208 rs4953273  WIN163  13  28.66  32.62 0.2553 0.6133 rs10865208 rs4953273  WIN164  33  39.23  24.81 3.248 0.0715 rs4953273  rs10189339  WIN164  31  42.39  39.94 0.07288 0.7872 rs4953273  rs10189339  WIN164  21  12.49  30.55 7.582 0.0059 rs4953273  rs10189339  WIN169  44  10.83  25.52 5.941 0.0148 rs4953288  rs13036100  WIN169  24  20.04  18.65 0.04972 0.8236 rs4953288  rs13036100  WIN169  42  30.61  18.66 2.902 0.0885 rs4953288  rs13036100  WIN169  22  32.56  31.21 0.02857 0.8658 rs4953288  rs13036100      123  APPENDIX3 Table 15 - 2-SNP haplotype TDT p-values for all haplotypes in a window when one of the two markers was associated with PCD in the single marker association tests for the 4q12 region. A,C,G,T alleles are represented by  1,2,3,4. P values <0.05 are in bold. T= number of transmitted haplotypes, U= number of untransmitted haplotypes  Window  Haplotype  T  U  CHISQ  P  SNPs  WIN4  12  6.21  5.38 0.05939 0.8075 rs4864596  rs10222715  WIN4  32  25.62  37.63 2.28 0.1310 rs4864596  rs10222715  WIN4  11  33.86  20.38 3.349 0.0672 rs4864596  rs10222715  WIN4  31  39.97  42.27 0.06431 0.7998 rs4864596  rs10222715  WIN42  44  1.852  2.426 0.07694 0.7815 rs17425986 rs9999844  WIN42  24  31.3  24.15 0.9214 0.3371 rs17425986 rs9999844  WIN42  42  4.148  12.3 4.037 0.0445 rs17425986 rs9999844  WIN42  22  32.43  30.85 0.03914 0.8432 rs17425986 rs9999844  WIN64  12  14  9 1.087 0.2971 rs7683304  rs4449471  WIN64  22  42  25 4.313 0.0378 rs7683304  rs4449471  WIN64  21  22  44 7.333 0.0068 rs7683304  rs4449471  WIN65  23  25  28 0.1698 0.6803 rs4449471  rs6838712  WIN65  21  45  22 7.896 0.0050 rs4449471  rs6838712  WIN65  13  24  44 5.882 0.0153 rs4449471  rs6838712  WIN66  31  25  48 7.247 0.0071 rs6838712  rs4865252  WIN66  13  42  22 6.25 0.0124 rs6838712  rs4865252  WIN66  11  16  13 0.3103 0.5775 rs6838712  rs4865252  WIN67  11  21  39 5.4 0.0201 rs4865252  rs4865258  WIN67  33  21  11 3.125 0.0771 rs4865252  rs4865258  WIN67  31  26  18 1.455 0.2278 rs4865252  rs4865258  WIN72  21  45  26 5.085 0.0241 rs17088743 rs1509541  WIN72  34  6  11 1.471 0.2253 rs17088743 rs1509541  WIN72  24  24  38 3.161 0.0754 rs17088743 rs1509541  WIN141  42  24  37 2.77 0.0960 rs925421  rs1397814  WIN141  22  14  22 1.778 0.1824 rs925421  rs1397814  WIN141  41  1  2 0.3333 0.5637 rs925421  rs1397814  WIN141  21  52  30 5.902 0.0151 rs925421  rs1397814  WIN142  14  29  37 0.9697 0.3248 rs1397814  rs1858317  WIN142  12  56  29 8.576 0.0034 rs1397814  rs1858317  WIN142  22  31  50 4.457 0.0348 rs1397814  rs1858317  WIN147  33  11  8 0.4737 0.4913 rs17238254 rs1513537  WIN147  23  43  27 3.657 0.0558 rs17238254 rs1513537  WIN147  24  28  47 4.813 0.0282 rs17238254 rs1513537   124   Table 15 – continued  Window  Haplotype  T  U  CHISQ  P  SNPs  WIN148  34  12  11 0.04348 0.8348 rs1513537  rs17089687  WIN148  42  28  47 4.813 0.0282 rs1513537  rs17089687  WIN148  32  43  25 4.765 0.0291 rs1513537  rs17089687  WIN180  33  20  37 5.07 0.0243 rs7690448  rs1511104  WIN180  31  25  8 8.758 0.0031 rs7690448  rs1511104  WIN180  11  19  19 0 1.0000 rs7690448  rs1511104  WIN181  31  24.45  37.08 2.589 0.1076 rs1511104  rs11929881  WIN181  11  25.51  11.68 5.149 0.0233 rs1511104  rs11929881  WIN181  13  12.24  10.29 0.1683 0.6816 rs1511104  rs11929881  WIN181  33  8.838  12 0.4799 0.4885 rs1511104  rs11929881  WIN182  34  0.8392  0 0.8392 0.3596 rs11929881 rs7690173  WIN182  32  20.9  21.64 0.01291 0.9095 rs11929881 rs7690173  WIN182  12  22.42  36 3.157 0.0756 rs11929881 rs7690173  WIN182  14  21.22  7.741 6.276 0.0122 rs11929881 rs7690173  WIN183  21  12  11 0.04348 0.8348 rs7690173  rs10021608  WIN183  41  23  9 6.125 0.0133 rs7690173  rs10021608  WIN183  23  14  29 5.233 0.0222 rs7690173  rs10021608  WIN184  11  24  14 2.632 0.1048 rs10021608 rs13141378  WIN184  13  10  5 1.667 0.1967 rs10021608 rs13141378  WIN184  33  15  30 5 0.0254 rs10021608 rs13141378  WIN185  14  11.46  3.418 4.35 0.0370 rs13141378 rs4469109  WIN185  34  3.372  10.01 3.296 0.0695 rs13141378 rs4469109  WIN185  12  16.25  11.89 0.6746 0.4114 rs13141378 rs4469109  WIN185  32  23.94  29.7 0.6188 0.4315 rs13141378 rs4469109  WIN195  44  11  23 4.235 0.0396 rs10033568 rs10517515  WIN195  14  31  17 4.083 0.0433 rs10033568 rs10517515  WIN195  11  8  10 0.2222 0.6374 rs10033568 rs10517515  WIN209  13  5.085  3.406 0.3322 0.5644 rs2880143  rs10517524  WIN209  33  29.91  11.59 8.086 0.0045 rs2880143  rs10517524  WIN209  11  29.69  33.14 0.1898 0.6631 rs2880143  rs10517524  WIN209  31  32.86  49.41 3.328 0.0681 rs2880143  rs10517524  WIN210  14  31  31 0 1.0000 rs10517524 rs11131298  WIN210  32  37  17 7.407 0.0065 rs10517524 rs11131298  WIN210  12  28  48 5.263 0.0218 rs10517524 rs11131298  WIN219  13  5.409  4.76 0.04139 0.8388 rs11131325 rs17238975  WIN219  33  41.57  26.25 3.461 0.0628 rs11131325 rs17238975  WIN219  11  29.07  39.45 1.571 0.2101 rs11131325 rs17238975  WIN219  31  35.41  41 0.4097 0.5221 rs11131325 rs17238975    125   Table 15 – continued  Window  Haplotype  T  U  CHISQ  P  SNPs  WIN220  33  47.55  26.64 5.894 0.0152 rs17238975 WIN220  WIN220  13  21.51  39.26 5.18 0.0229 rs17238975 WIN220  WIN220  31  4.687  6.945 0.4384 0.5079 rs17238975 WIN220  WIN220  11  32.64  33.55 0.01254 0.9108 rs17238975 WIN220     126  APPENDIX4 Table 166 - 2-SNP haplotype TDT p-values for the rs16888748|rs3804320 haplotypes in the 6p22.2 region. A,C,G,T alleles are represented by  1,2,3,4. P values <0.05 are in bold. T= number of transmitted haplotypes, U= number of untransmitted haplotypes  Window  Haplotype  T  U  CHISQ  P  SNPs  WIN5  24  1  8  5.444 0.01963 rs16888748 rs3804320  WIN5  44  20  17  0.2432 0.6219 rs16888748 rs3804320  WIN5  23  25  20  0.5556 0.4561 rs16888748 rs3804320  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items