Whole-exome sequencing in UK Biobank reveals rare genetic architecture for depression

Abstract

Nearly two hundred common-variant depression risk loci have been identified by genome-wide association studies (GWAS). However, the impact of rare coding variants on depression remains poorly understood. Here, we present whole-exome sequencing analyses of depression with seven different definitions based on survey, questionnaire, and electronic health records in 320,356 UK Biobank participants. We showed that the burden of rare damaging coding variants in loss-of-function intolerant genes is significantly associated with risk of depression with various definitions. We compared the rare and common genetic architecture across depression definitions by genetic correlation and showed different genetic relationships between definitions across common and rare variants. In addition, we demonstrated that the effects of rare damaging coding variant burden and polygenic risk score on depression risk are additive. The gene set burden analyses revealed overlapping rare genetic variant components with developmental disorder, autism, and schizophrenia. Our study provides insights into the contribution of rare coding variants, separately and in conjunction with common variants, on depression with various definitions and their genetic relationships with neurodevelopmental disorders.

Fig. 1. The association between exome-wide rare coding variant burdens and depression with seven different definitions.

Y-axis is the odds ratio (OR) of the association between rare variant burden and depression risk. Protein-coding genes were stratified by gene Loss-of-Function (LoF) intolerant with pLI score into (a) pLI ≥ 0.9 (LoF intolerant) and (b) pLI < 0.9 (LoF tolerant). Rare variants were grouped by functional impact from the most to least severe: protein-truncating, missense (MPC > 2, 2 ≥ MPC > 1, 1 ≥ MPC > 0), other missense (missense variants without MPC score annotation) and synonymous variants. Missense variants in genes (pLI < 0.9) were only annotated into two categories, 2 ≥ MPC > 1 and 1 ≥ MPC > 0. The sample size for each depression definition are as follows: GPpsy: N_cases = 111,712, N_controls = 206,617; Psypsy: N_cases = 36,556, N_controls = 282,452; DepAll: N_cases = 20,547, N_controls = 55,746; SelfRepDep: N_cases = 20,120, N_controls = 226,578; EHR: N_cases = 10,449, N_controls = 246,719; lifetimeMD: N_cases = 15,580, N_controls = 43,104; MDDRecur: N_cases = 9462, N_controls = 43,104. The gray dashed line represents the null (OR = 1). Each point shows the point estimate of OR from logistic regression. Bars show 95% confidence intervals (CI). *Odds ratios with significant p based on Bonferroni-adjusted significance threshold p < 4.20 × 10⁻⁴ = 0.05/119 (two-sided Wald test; Supplementary Data 3).

Fig. 2. Genetic correlations estimated from rare genetic burden and common variants across depression definitions.

Pairwise burden genetic correlation of depression definitions estimated from (a) PTV (MAF < 0.01) and (b) missense variants (MAF < 0.01). c Pairwise genetic correlations (r_G) estimated from common variants between depression definitions (from Cai et al.). All pairwise genetic correlation estimates were significant at Bonferroni-adjusted threshold (p < 7.94 × 10⁻⁴ = 0.05/63) based on two-sided Wald tests (Supplementary Data 9), except for the burden genetic correlation from PTV between DepAll and lifetimeMDD (p = 8.7 × 10⁻³) and between DepAll and MDDRecur (p = 3.77 × 10⁻³).

Fig. 3. Additive contributions from rare and common variants to depression risk.

The prevalence of (a) EHR-, (b) Psypsy- and (c) DepAll-defined depression against PRS percentile, stratified by exome-wide PTV or damaging missense variant carrier status. The lines represent the locally fitted regression line by LOESS regression, and the gray shading corresponds to the 95% confidence interval of the fitted regression. The sample sizes for each depression definition are as follows: Psypsy: N_cases = 36,556, N_controls = 282,452; DepAll: N_cases = 20,547, N_controls = 55,746; EHR: N_cases = 10,449, N_controls = 246,719.

Fig. 4. The effects of rare coding variants in psychiatric and neurodevelopmental disease genes on EHR-defined depression.

a The effect of rare variants in psychiatric and neurodevelopmental disease associated genes from previously published exome studies and (b) genes from previously published GWAS for psychiatric disorders. We aggregated rare variants of each type (PTV, missense and synonymous) on 8 disease gene sets. From exome studies, we identified 102 autism (ASD) genes (FDR < 0.1), 285 developmental disorder (DD/ID) genes (Bonferroni significant), and 32 schizophrenia (SCZ) genes (FDR < 0.05). From GWAS, we identified 269 genes for depression (MDD), 218 genes for bipolar disorder (BP), and 3542 complete positionally mapped genes (“SCZ GWAS complete”), 114 prioritized protein-coding genes (“SCZ GWAS prioritized”) and 69 fine-mapped genes (“SCZ GWAS fine-mapped”) for schizophrenia. Y-axis is the odds ratio of the association between rare variant burden for each gene set and depression risk. The gray dashed line represents the null (OR = 1). Each point shows the odds ratio from logistic regression. Bars show 95% confidence intervals. *Odds ratios with significant p based on Bonferroni-adjusted significance threshold p < 1.04 × 10⁻³ = 0.05/48 (two-sided Wald test; Supplementary Data 13).