Identifying novel genetic loci associated with polycystic ovary syndrome based on its shared genetic architecture with type 2 diabetes

Abstract

Genome-wide association studies (GWAS) have identified several common variants associated with polycystic ovary syndrome (PCOS). However, the etiology behind PCOS remains incomplete. Available evidence suggests a potential genetic correlation between PCOS and type 2 diabetes (T2D). The publicly available data may provide an opportunity to enhance the understanding of the PCOS etiology. Here, we quantified the polygenic overlap between PCOS and T2D using summary statistics of PCOS and T2D and then identified the novel genetic variants associated with PCOS behind this phenotypic association. A bivariate causal mixture model (MiXeR model) found a moderate genetic overlap between PCOS and T2D (Dice coefficient = 44.1% and after adjusting for body mass index, 32.1%). The conditional/conjunctional false discovery rate method identified 11 potential risk variants of PCOS conditional on associations with T2D, 9 of which were novel and 6 of which were jointly associated with two phenotypes. The functional annotation of these genetic variants supports a significant role for genes involved in lipid metabolism, immune response, and the insulin signaling pathway. An expression quantitative trait locus functionality analysis successfully repeated that 5 loci were significantly associated with the expression of candidate genes in many tissues, including the whole blood, subcutaneous adipose, adrenal gland, and cerebellum. We found that SCN2A gene is co-localized with PCOS in subcutaneous adipose using GWAS-eQTL co-localization analyses. A total of 11 candidate genes were differentially expressed in multiple tissues of the PCOS samples. These findings provide a new understanding of the shared genetic architecture between PCOS and T2D and the underlying molecular genetic mechanism of PCOS.