Sprache, Kommunikation und soziale Entwicklung

according to the manufacturer’s instructions or the KBiosciences service using their in-house technology. Both quantitative and case-control analyses were performed within PLINK (1.07) applying additive tests of association.
Results
    Candidate genes for dyslexia (
MRPL19/C2ORF3
locus,
KIAA0319
and
DCDC2
) and SLI (
CMIP
and
ATP2C2
) were tested for association with reading and language-related traits in a general population sample derived from the ALSPAC cohort (N = 3725) after filtering for missing data, ethnicity and low IQ. The exclusion of children with a performance IQ < 85 was aimed at removing confounding effects where children would perform low on reading and language tasks because of different underlying problems. Significant associations were observed for markers within
CMIP, DCDC2
and
KIAA0319
with test of single-world reading specifically but not with other reading or language related traits. To further explore whether the associations were genuinely with general reading abilities or rather might have been driven by the most impaired individuals on the reading tasks the sample was stratified. Therefore, a subgroup of children meeting criteria for a dyslexia diagnosis was identified and then excluded from association analysis. While
KIAA0319
and
CMIP
showed the same degree of association, the signals for
DCDC2
tended to disappear suggesting that a small proportion of individuals with the most severe phenotype were indeed at the basis of the original observation in the larger sample. These results suggest that different processes may control reading abilities.
KIAA0319
and
CMIP
seem to be associated with general reading skills, while
DCDC2
seems tobe specifically associated with dyslexia. This hypothesis was supported by a case-control analysis where the cases were represented by the same subgroups of children removed from the quantitative analysis. Super-controls were selected among the children scoring very high on the reading tests.
DCDC2
showed the strongest signal supporting the hypothesis of a specific role in dyslexia. The same case-control approach led also to interesting observations on the effects of comorbidity. When the group of cases included children that met criteria for both dyslexia and SLI diagnosis the
DCDC2
association was stronger. These results suggest the importance of including children with signs of comorbidity for SLI or ADHD when designing association studies of dyslexia.
    In a different study ALSPAC was used as a replication sample following the identification of
PCSK6
through a genome-wide association study (GWAS) conducted in children with dyslexia (Scerri et al. 2011).
PCSK6
is a very interesting candidate for handedness because of its known interaction with
NODAL
in establishing left-right body axis determination in early development. ALSPAC is one of the few available samples characterised with the peg-board test. To match the phenotypic characteristics of the discovery sample, a subgroup of individuals with severe reading impairment was selected for replication analysis. The same
PCSK6
genetic markers yielded a significant association with handedness in this subgroup providing a robust replication. Meta-analysis across samples reached genome-wide significance level (
P
< 5 × 10 – 7 ). To assess this result in the general population, analysis of
PCSK6
in the remaining of the ALSPAC children (N = 2667), showed a significant association but with an opposite allelic trend. These results suggest a specific effect of this association in people with dyslexia. It is important to note that an association between handedness and dyslexia was not observed but rather that the effects of these genetic markers vary in different groups of people possibly because of interaction with a different genetic background.
Conclusions
    Genetic analysis in population-based cohorts has proved to be a useful tool to further dissect at phenotype level associations with language-related disorders. First of all, such cohorts represent valid tools to replicate associations identified in patient cohorts. The observation that genetic variants associated with dyslexiado influence reading abilities in the normal range of variation supports robustly to the idea that dyslexia can be considered the lower tail of the reading ability distribution across the population rather than a categorically defined condition. The combination of large sample sizes and a wide range of phenotypic measures available for the