Abstract
OBJECTIVE: A previous analysis of 200,000 exome-sequenced UK Biobank participants using weighted burden analysis of rare, damaging variants failed to identify any genes associated with risk of affective disorder requiring specialist treatment. Exome-sequence data has now been made available for the remaining 270,000 participants and a two-stage process was applied in order to test for association in this second sample using only genes showing suggestive evidence for association in the first sample. METHODS: Cases were defined as participants who reported having seen a psychiatrist for 'nerves, anxiety, tension or depression'. Exhaustive testing of the first sample was carried out using rare variant analyses informed by 45 different predictors of impact of nonsynonymous variants. The 100 genes showing the strongest evidence for association were then analysed in the second sample using the same predictor as had been most statistically significant in the first sample. RESULTS: The results for the 100 nominated genes conformed closely with the null hypothesis, with none approaching statistical significance after correction for multiple testing. CONCLUSION: Risk of common affective disorder, even if severe enough to warrant specialist referral, is not sufficiently impacted by effects of rare variants in a small enough number of genes that effects can be detected even with large sample sizes. Actionable results might be obtained with a more extreme phenotype but very significant resources would be required to achieve adequate power. This research has been conducted using the UK Biobank Resource.