Shared Genetic Variants in Speech and Language Traits

Introduction            Oral and written language abilities during childhood are related to educational attainment, occupational success and health during later life. Many of these traits are highly heritable... [ view full abstract ]

Introduction

           Oral and written language abilities during childhood are related to educational attainment, occupational success and health during later life. Many of these traits are highly heritable and share genetic links with verbal intelligence quotient (VIQ) scores. The nature of this shared genetic basis is, however, not well understood. We aim to identify and characterise the proportion of genetic variation in the genome, as tagged by common markers, that can explain phenotypic variance across multiple oral and written language abilities during childhood.

Method

We studied 13 traits (comprehension, reasoning, reading, spelling, phonological awareness/memory abilities) in 7 to 12 years-old children from the Avon Longitudinal Study of Parents and Children (N≤5,919). The traits were heritable (SNP-h² from 0.25(S.E.=0.07) to 0.49(S.E.=0.07)) and genetically correlated with VIQ scores (r_g from 0.62(S.E.=0.10) to 0.89(S.E.=0.07)). We carried out a cross-phenotype meta-analysis of correlated traits based on GWAS summary statistics to identify genetic variants that share similar effects across traits (core-SNPs), restricting the analysis to HapMap3 SNPs.

Results

            Tagging ~10% of the genome, the core-SNPs explained a large proportion of SNP-h² per trait (minimum enrichment 11-fold(S.E.=0.22), P=1×10^-5), although these results might be inflated because of over-fitting. Evidence for enrichment was also observed for performance IQ scores (10-fold(S.E.=0.21), P=8×10^-6), but not for height (P=0.06) or gestational age (P=0.32). An analysis of core-SNPs in the UK Biobank (N=36,035, 40-69 years) also showed some evidence of enrichment with respect to fluid IQ scores (2-fold(S.E.=0.03), P=0.004) and its verbal subset (2-fold(S.E.=0.03), P=0.02).

Discussion

            Our analyses suggest that a defined set of variants in the genome can explain a large proportion of phenotypic variance across multiple traits that require directly and indirectly the mastering of language skills. However, genetic variants contributing to phenotypic variation during childhood may differ from those in adulthood, and our findings require further replication in age-matched population-based cohorts.