Objective: Dyslexia-susceptibility-1-candidate-1 (DYX1C1) was the first gene associated with dyslexia. Since the original report of 2003, eight replication attempts have been published reporting discordant results. As the dyslexia community still considers the role of DYX1C1 unsettled, we explored the contribution of this gene in a sample of 366 trios of German descent.
Methods: To the common four markers used in previous studies, we added two new single nucleotide polymorphisms found by resequencing both the putative regulatory and coding region of the gene in randomly selected cases and controls. As linkage disequilibrium blocks of the region were not easy to define, we approached the association problem by running a transmission disequilibrium test over sliding windows of dimension 1 to 6 on consecutive markers. The significance of this test was calculated generating the empirical distribution of the global P value by simulating the data. As our study sample had a large female proband content, we also stratified our analysis by sex.
Results: We found statistically significant association with global corrected P value of 0.036. The three-marker haplotype G/G/G spanning rs3743205/rs3743204/rs600753 was most associated with a P value of 0.006 and odds ratio 3.7 (95% confidence interval: 1.4–9.6) in female probands. A detailed haplotype–phenotype analysis revealed that the dyslexia subphenotype short-term memory contributed mainly to the observed findings. This is in accordance with a recent short-term memory–DYX1C1 association in an independent sample of dyslexia.
Conclusion: As significant association was proved in our sample, we could also conclude that denser maps, sex information, and well-defined subphenotypes are crucial to correctly determine the contribution of DYX1C1 to dyslexia.
aInstitute of Human Genetics
bDepartment of Genomics, Life and Brain Center, University of Bonn, Bonn
cDepartment of Child and Adolescent Psychiatry, Psychosomatics and Psychotherapy, University of Munich
dMax Planck Institute of Psychiatry, Munich
eDepartment of Child and Adolescent Psychiatry and Psychotherapy, University of Würzburg, Würzburg
fDepartment of Child and Adolescent Psychiatry and Psychotherapy, University of Marburg, Marburg
gInstitute for Medical Biometry and Statistics, University of Lübeck, Lübeck, Germany
hDepartment of Medical Genetics, University of Helsinki, Helsinki, Finland
Departments of iClinical Research Center
jBiosciences and Nutrition, Karolinska Institutet, Huddinge, Sweden
kUnit on the Genetic Basis of Mood and Anxiety Disorders, National Institute of Mental Health, National Institutes of Health, Bethesda, USA
Correspondence to Professor Juha Kere, Department of Biosciences and Nutrition, Karolinska Institutet, 14157 Huddinge, Sweden
Tel: +46 8 6089 158, fax: +46 8 7745 538;
Received 9 July 2008 Revised 25 August 2008 Accepted 25 August 2008
Dyslexia (MIM 12770) affects 5–12% of school-age children and is one of the most common neurobehavioral disorders (Katusic et al., 2001). In recent years, linkage studies have identified regions likely to harbor genes contributing to dyslexia (Schumacher et al., 2007). In particular, nine chromosomal regions, DYX1 through DYX9, have been reported in dyslexia and are listed by the Human Gene Nomenclature Committee. Close to one of these loci, the gene dyslexia-susceptibility-1-candidate-1 (DYX1C1, MIM 608706) on chromosome 15q21 was identified as a susceptibility factor for dyslexia. A chromosomal translocation t(2;15)(q11;q21) causing a disruption of DYX1C1 cosegregated with dyslexia in a two-generation family, and two genetic variants at the DYX1C1 locus (rs3743205 and 1249G>T; submitted to database single nucleotide polymorphism (SNP) and designated rs61761345) showed significant association with dyslexia in an independent sample of Finnish origin (Taipale et al., 2003). Positive association findings at the DYX1C1 locus were subsequently reported by three independent groups. Wigg et al. (2004) studied both markers analyzed by Taipale et al. (2003) and found significant association in a dyslexia sample of European–Canadian descent. Marino et al. (2007) observed, in an Italian dyslexia sample, association at both variants using short-term memory (STM) as a phenotypic trait. It is well established that STM plays an important role in the development of dyslexia (Howes et al., 1999). No significant association was, however, observed in this sample using a categorical diagnosis of dyslexia (Marino et al., 2005). The authors claimed that this discrepancy could be because of the fact that in their dyslexia study unaffected siblings were not included, reducing the overall power of the association test. Most recently, Bates et al. (2007) reported on significant association between DYX1C1 markers and dyslexia in 789 families comprising twin series from Australia. The reported association evidence is, however, complicated by the fact that different alleles and/or haplotypes were associated across studies. Along this line, although Scerri et al. (2004) found significant association at the two-marker haplotype, rs3743205–rs61761345, in a UK dyslexia sample, they interpreted their results as negative replication. Three other studies failed to find any DYX1C1 association in their dyslexia samples, which were mainly of European descent (Bellini et al., 2005; Cope et al., 2005; Marino et al., 2005; Meng et al., 2005).
In the light of the cited studies, the role of DYX1C1 in dyslexia is thus considered unsettled (Fisher and Francks, 2006). In this study, we aimed to explore the contribution of DYX1C1 to dyslexia in the German population and analyzed a large family-based sample of 366 trios. In a first step, we looked for association using the most significant associated SNPs reported by Taipale et al. (2003). Although we failed to detect association using these two markers we extended our analysis using a combination of resequencing the entire susceptibility locus and further SNP genotyping for a better DYX1C1 marker coverage.
Materials and methods
A total of 366 family trios of German descent (300 male and 66 female indices, see supplementary Table 1) were used for this study, which represent a part of a large ongoing national recruitment effort for families with dyslexia (Schumacher et al., 2006; Schulte-Körne et al., 2007). All individuals, and in cases of children younger than 14 years their parents, gave written informed consent to participation in the study. The families were recruited from the Departments of Child and Adolescent Psychiatry and Psychotherapy at the Universities of Marburg and Würzburg, and the study was approved by the local ethics committees.
The diagnostic inclusion criteria and phenotypic measures have been described in detail previously (Schulte-Körne et al., 1996; Schulte-Körne et al., 2001; Ziegler et al., 2005; Schumacher et al., 2006; Schulte-Körne et al., 2007). Briefly, the diagnosis of dyslexia was based on the spelling score using the T distribution of the general population. To be diagnosed as dyslexic, the child had to meet the following discrepancy criterion: based on the correlation between IQ and spelling of 0.4 (Schulte-Körne et al., 2001), an anticipated spelling score was calculated. The child was classified as dyslexic if the discrepancy between the anticipated and the observed spelling scores was at least 1 standard deviation (1 SD). Probands and all siblings fulfilling the inclusion criteria were assessed with several psychometric tests. These tests targeted different aspects of the dyslexia, that is, word reading, phonological awareness, phonological decoding, rapid naming, STM, and orthographic coding (see supplementary information for a more detailed description). As it has been previously found that association findings might become stronger in samples of severely affected individuals (Schumacher et al., 2006), we also stratified our data by the severity of the phenotype (see supplementary information and Schumacher et al., 2006 for the definition of disease severity). As affected probands were defined by discrepancy to the observed spelling score of at least ≥1 SD, two subgroups of families of more-severely affected children were defined by probands discrepancy of at least ≥2 and ≥2.5 SD, respectively.
Single nucleotide polymorphism genotyping
SNP genotyping was carried out using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (Sequenom, San Diego, California, USA). Extension products were analyzed by a Mass ARRAY mass spectrometer (Bruker Daltonik, Bremen, Germany) and peaks were identified using the SpectroTYPER software (Sequenom). All genotypes were scored independently by two individuals blind to the disease status, and tested for Mendelian inheritance using PedCheck. Genotype success ratio was greater than 0.80 for all analyzed SNPs. Detailed information on primer sequences, PCR amplification, genotyping procedure, and genotype calling can be obtained on request.
All PCR products were cleaned from unincorporated primers and deoxynucleotide triphosphates using shrimp alkaline phosphatase and exonuclease I and further sequenced using a DYEnamic ET Dye terminator kit (Amersham Biosciences, Piscataway, New Jersey, USA). Sequencing products were electrophoresed using a MegaBACE 1000 instrument and MegaBACE long read matrix; visualized using the Sequence Analyzer v3.0 software (Amersham Biosciences) and further aligned using the Pregap and Gap4 software (Staden package; www.staden.sourceforge.net). In addition, a separate viewer compared each FASTA output from sequencing results to corresponding genomic sequences (NT_010194) using BLAST 2 sequences. Detailed information on PCR amplification, sequencing, and analysis is available on request.
Association between dyslexia and SNPs and haplotypes was assessed using the software PDTPHASE (Dudbridge, 2003). As there is evidence for a sex-specific genetic influence on dyslexia (Rutter et al., 2004; Harlaar et al., 2005), tests were performed for all families as well as separately for families with a female and a male index patient, respectively. Haplotype analysis was performed using a sliding window approach, with window sizes ranging from 1 to 6. For each sliding window, differences in the haplotype distribution between affected individuals and parents (global P values) as well as transmission rates of each individual haplotype were analyzed. To assess the significance of our findings, the empirical distribution of the test statistics was generated by using 10 000 pedigrees/genotypes simulations under the null hypothesis using SimPed (Leal et al., 2005). To adjust the nominal P values for the fact of having tested male and female probands separately as well as jointly, a Westfall–Young permutation-based minimum P step-down procedure was done based on the 10 000 permutations performed (Westfall and Young, 1993). Finally, our testing strategy provided a global empirical association P value for the whole sample based on 21×3 tests (six single SNPs and 15 haplotypic tests on males, females, and the joint sample, respectively). The analysis of quantitative phenotypes was performed using QPDTPHASE, which is a quantitative trait implementation of the PDT. We used the UNPHASED implementation as for the PDT (Dudbridge, 2003).
Given the previous inconsistent association findings, we aimed to explore the contribution of the DYX1C1 locus to dyslexia in a large family sample of German descent (366 trios). To increase the DYX1C1 marker coverage and the resolution of the locus-specific haplotypic structure, we resequenced the entire coding DYX1C1 region (10 exons, including the flanking sequences) as well as the corresponding 5′ and 3′ untranslated regions in 10 patients and 10 controls of German descent, all randomly selected. This allows the detection of SNPs with a minor allele frequency of greater than 0.10 with a power of ≥98% (Gregorius, 1980). Comparison of the analyzed gene region to the public sequence (NT_010194) revealed three previously unstudied SNPs within the DYX1C1 promoter region – rs12899331, rs16787, and rs8043269. On the basis of resequencing results and on information from the public database single nucleotide polymorphism we then selected for further genotyping the three identified promoter variants, two additional SNP markers located within the coding gene region, rs3743204 and rs600753 and the two SNPs originally found associated in dyslexia, namely rs3743205 and rs61761345 (Fig. 1). Of these variants rs8043269 was discarded as it failed our assay design criteria.
By running our algorithm, we found significant association in our sample with global P value of 0.036. This P value is properly corrected for multiple testing and sex stratification. A posteriori inspection of the individual sliding window tests provided the best association finding in the female's group for the three-marker haplotype G/G/G at rs3743205/rs3743204/rs600753. This carried an odds ratio of 3.71 (95% confidence interval: 1.44–9.60, P value=0.006). The frequency of this common haplotype was estimated to be 0.49 in female patients (Table 1). A replication of the whole analysis over the severity-stratified sample did not provide any improvement of the association results.
Further, we sought whether the observed association was mainly driven by a particular subphenotype of dyslexia. The quantitative dyslexia subphenotypes spelling, word reading, phonological decoding, phonological awareness, orthographic processing, rapid naming, and STM were tested using QPDTPHASE. The same common three-marker haplotype G/G/G at rs3743205/rs3743204/rs600753, which was associated in our entire female sample, showed significant results by using STM as a phenotypic trait (global P value of 0.011, haplotype-based P value of 0.0046). All other subphenotypes showed no association to this particular haplotype in females and the haplotype–STM association was not significant in the entire or male samples (data not shown). This result is consistent with recently reported findings of association of dyslexia to STM.
In this study, we aimed to explore the contribution of DYX1C1 to dyslexia in the German population. We resequenced the coding and putative regulatory gene region in randomly selected patients and controls and carried out association studies using genetic variants covering the DYX1C1 locus in a large family-based sample. For the following reasons we believe that our results might be of interest for the dyslexia research community: First, we identified a three-marker DYX1C1 haplotype, which was significantly associated to female dyslexia patients. This finding still remains significant after correction for multiple testing and was associated with an odds ratio of 3.71. Although we are aware of the fact that our female trios represent only a small fraction of our entire dyslexia sample (66 females of 366 cases), our results indicate that consideration of sex can influence the outcome of association findings in dyslexia. To our knowledge, this study represents the first molecular genetic study of dyslexia applying a sex-separated analysis, although differences in prevalence rates between females and males have been reported in epidemiological dyslexia studies (Rutter et al., 2004) and sex-specific differences have already been reported for a variety of other complex traits in humans (Rutter et al., 2004; Weiss et al., 2006). Moreover, our findings may shed some light on the contradictory DYX1C1 dyslexia association findings reported so far. As the female-to-male probands ratios vary largely across all previously analyzed samples, the statistical power in the detection of a DYX1C1 association may also vary across the different samples.
Second, in our sample of female patients we observed a significant association of the dyslexia subphenotype STM by applying an exploratory haplotype–phenotype analysis. Although this finding needs further replication by independent studies, our results are in accordance with the observed STM–DYX1C1 association in the dyslexia sample of Marino et al. (2007). Both studies point to STM as a dyslexia subphenotype, which might be more directly influenced by the genetic variation at the DYX1C1 locus than the disorder itself. Third, our results together with the study of Bates et al. (2007) indicate that genotyping the SNP markers, which showed association in the Finnish population (Taipale et al., 2003), is insufficient in capturing the dyslexia risk haplotype in Central European populations. In both studies the detection of association was only possible by increasing the DYX1C1 marker coverage. Finally, our results and the findings of other studies (Scerri et al., 2004; Wigg et al., 2004) indicate that the putative DYX1C1-causing mutation in Central Europeans is located on the common haplotype – G/G – at rs3743205–rs61761345. This might point to a common ancestor in Central Europeans, whereas the results in the Finnish and the Italian populations point to independent DYX1C1 mutation events (Taipale et al., 2003; Marino, Citterio et al., 2007).
We are aware of the fact that this study is hampered by several limitations, for example, moderate number of female patients, exploratory analysis of dyslexia subphenotypes, and moderate DYX1C1 marker coverage. Our findings are, however, in line with five independent reports indicating a role for DYX1C1 in the development of dyslexia or dyslexia-related phenotypes (Taipale et al., 2003; Scerri et al., 2004; Wigg et al., 2004; Bates et al., 2007; Marino, Citterio et al., 2007). On the basis of this association evidence, it seems premature to reject DYX1C1 from the list of potential candidate genes in dyslexia. Instead, it seems necessary to carry out more detailed DYX1C1–dyslexia association studies in future. Those analyses should include large dyslexia samples from different populations, the consideration of sex and detailed subphenotype information as well as maximum marker coverage across the gene locus. The results of the forthcoming genome-wide association studies (e.g. www.neurodys.com) may allow the field to determine the relevance of DYX1C1 to dyslexia.
Supplementary data are available at The Psychiatric Genetics Journal Online (http://www.psychgenetics.com).
The authors are grateful to all investigated individuals for their cooperation in this study. This study was supported by the Deutsche Forschungsgemeinschaft (DFG). M.M.N. received support for this study from the Alfried Krupp von Bohlen und Halbach-Stiftung; J.S. is a research fellow of the NIH/DFG Research Career Transition Awards Program; M.P.J. is a recipient of a research position from Swedish Research Council; I.T.P. is supported by a grant from Swedish Brain Foundation; H.A. is supported by the Centennial Foundation of Helsingin Sanomat; M.Z. is partly supported by the Bioinformatics and Expression Analysis Core Facility at Karolinska Institutet; and J.K. is supported by grants from Swedish Research Council, Academy of Finland, Sigrid Jusélius Foundation, and Päivikki and Sakari Sohlberg Foundation.
Bates T, Luciano M, Lind P, Wright M, Montgomery G, Martin N (2007). Genetic influences on dyslexia: support for DYX1C1 and DCDC2. World Congress on Psychiatric Genetics XII, 7–11 October 2007, New York City.
Bellini G, Bravaccio C, Calamoneri F, Donatella Cocuzza M, Fiorillo P, Gagliano A, et al. 2005. No evidence for association between dyslexia and DYX1C1 functional variants in a group of children and adolescents from Southern Italy. J Mol Neurosci 27:311–314.
Cope NA, Hill G, Van Den Bree M, Harold D, Moskvina V, Green EK, et al. 2005. No support for association between dyslexia susceptibility 1 candidate 1 and developmental dyslexia. Mol Psychiatry 10:237–238.
Dudbridge F 2003. Pedigree disequilibrium tests for multilocus haplotypes. Genet Epidemiol 25:115–121.
Fisher SE, Francks C 2006. Genes, cognition and dyslexia: learning to read the genome. Trends Cogn Sci 10:250–257.
Gregorius HR 1980. The probability of losing an allele when diploid genotypes are sampled. Biometrics 36:643–652.
Harlaar N, Spinath FM, Dale PS, Plomin R. 2005. Genetic influences on early word recognition abilities and disabilities: a study of 7-year-old twins. J Child Psychol Psychiatry 46:373–384.
Howes NL, Bigler ED, Lawson JS, Burlingame GM. 1999. Reading disability subtypes and the test of memory and learning. Arch Clin Neuropsychol 14:317–339.
Katusic SK, Colligan RC, Barbaresi WJ, Schaid DJ, Jacobsen SJ. 2001. Incidence of reading disability in a population-based birth cohort, 1976–1982, Rochester, Minn. Mayo Clin Proc 76:1081–1092.
Leal SM, Yan K, Muller-Myhsok B. 2005. SimPed: a simulation program to generate haplotype and genotype data for pedigree structures. Hum Hered 60:119–122.
Marino C, Giorda R, Luisa Lorusso M, Vanzin L, Salandi N, Nobile M, et al. 2005. A family-based association study does not support DYX1C1
on 15q21.3 as a candidate gene in developmental dyslexia. Eur J Hum Genet 13:491–499.
Marino C, Citterio A, Giorda R, Facoetti A, Menozzi G, Vanzin L, et al. 2007. Association of short-term memory with a variant within DYX1C1 in developmental dyslexia. Genes Brain Behav 6:640–646.
Meng H, Hager K, Held M, Page GP, Olson RK, Pennington BF, et al. 2005. TDT-association analysis of EKN1 and dyslexia in a Colorado twin cohort. Hum Genet 118:87–90.
Rutter M, Caspi A, Fergusson D, Horwood LJ, Goodman R, Maughan B, et al. 2004. Sex differences in developmental reading disability: new findings from 4 epidemiological studies. JAMA 291:2007–2012.
Scerri TS, Fisher SE, Francks C, Macphie IL, Paracchini S, Richardson AJ, et al. 2004. Putative functional alleles of DYX1C1 are not associated with dyslexia susceptibility in a large sample of sibling pairs from the UK. J Med Genet 41:853–857.
Schulte-Körne G, Ziegler A, Deimel W, Schumacher J, Plume E, Bachmann C, et al. 1996. Familial aggregation of spelling disability. J Child Psychol Psychiatry 37:817–822.
Schulte-Körne G, Deimel W, Remschmidt H. 2001. Diagnosis of reading and spelling disorder. Z Kinder Jugendpsychiatr Psychother 29:113–116.
Schulte-Körne G, Ziegler A, Deimel W, Schumacher J, Plume E, Bachmann C, et al. 2007. Interrelationship and familiality of dyslexia related quantitative measures. Ann Hum Genet 71 (Pt 2):160–175.
Schumacher J, Anthoni H, Dahdouh F, König IR, Hillmer AM, Kluck N, et al. 2006. Strong genetic evidence of DCDC2 as a susceptibility gene for dyslexia. Am J Hum Genet 78:52–62.
Schumacher J, Hoffmann P, Schmal C, Schulte-Korne G, Nothen MM. 2007. Genetics of dyslexia: the evolving landscape. J Med Genet 44:289–297.
Taipale M, Kaminen N, Nopola-Hemmi J, Haltia T, Myllyluoma B, Lyytinen H, et al. 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.
Weiss LA, Pan L, Abney M, Ober C. 2006. The sex-specific genetic architecture of quantitative traits in humans. Nat Genet 38:218–222.
Westfall PH, Young SS 1993. Resampling-based multiple testing. New York: John Wiley & Sons, Inc.
Wigg KG, Couto JM, Feng Y, Anderson B, Cate-Carter TD, Macciardi F, et al. 2004. Support for EKN1 as the susceptibility locus for dyslexia on 15q21. Mol Psychiatry 9:1111–1121.
Ziegler A, König IR, Deimel W, Plume E, Nöthen MM, Propping P, et al. 2005. Developmental dyslexia – recurrence risk estimates from a german bi-center study using the single proband sib pair design. Hum Hered 59:136–143.