Abstracts

Rationale: Twin and family studies indicate approximately 50% of the variance in susceptibility to focal epilepsy can be attributed to genetic causes, but so far, genome-wide association studies (GWAS) have had limited success in identifying genetic susceptibility factors (i.e., the missing heritability problem).Methods: Using data from a genotyped cohort of UK patients with focal epilepsy, we apply heritability analysis to estimate the proportion of variance of susceptibility that can be explained by genotyped single nucleotide polymorphisms (SNPs). We also consider using heritability analysis for gene- and region-based tests of association, and for quantifying the importance of exons and rare variants. All our analyses can be performed using the freely-available software LDAK (http://dougspeed.com/ldak/).Results: Save for one promising region, we are unable to detect any individual loci which are significantly associated with susceptibility to partial epilepsy. By contrast, we find that the majority of heritability of partial epilepsy, about 80%, can be explained by the collective influence of genotyped common variants of weak effect. This result is confirmed by the ability to generate a significant risk prediction model. The gene and regional analyses indicate that this heritability is spread throughout the genome, and similarly throughout the allele frequency spectrum. Conclusions: Although these results demonstrate the ability of genotyped common variants to capture genetic effects on epilepsy susceptibility, they also imply that many hundreds of thousands of patients will be required to identify a reasonable share of causal variation through single-variant GWAS analysis. However, it remains that clinically useful prediction models can be achieved by recognising the many causal variants of weak effect, albeit that tens of thousands of individuals will be required.