Investigation of Recessive Effects in Schizophrenia Using Next-Generation Exome Sequence Data

A number of gene-wise approaches to analysis were applied to whole exome sequence data from 2545 Swedish schizophrenia cases with 2545 matched controls. A weighted burden test was used to detect dominant and additive effects. Recessive effects were investigated by testing whether there was an excess of cases bearing two or more rare, functional variants or whether there was an excess of cases in which both phased haplotypes carried at least one rare, functional variant. Counts for cases were compared with controls and also with the expectation under Hardy-Weinberg equilibrium. Analyses were performed using the SCOREASSOC program. No gene produced statistically significant results although COMT was highly ranked by the weighted burden test and within it the Ala72Ser polymorphism rs6267 had an uncorrected p value of 0.00003. A number of spurious results were generated, some apparently due to miscalling of homozygotes and others due to a failure to eradicate the effects of linkage disequilibrium between variants. These problems were not marked when using phased haplotypes but this method failed to produce any significant or suggestive findings. If there are exonic variants with recessive effects on the risk of schizophrenia, then the methods used were unable to detect them.