We performed an additional replication analysis of our two genome-wide significant SNPs using the publicly available data of the recently published GWAMA of depressive disorders in a sample of Chinese women (the CONVERGE study) (11). In CONVERGE, rs9825823 (odds ratio=1.01, P=0.12) and rs9323497 (odds ratio=0.97, P=0.0002, with a different direction of association than in our discovery sample) were not related to depression at the genome-wide significance level although the latter reached nominal significance. However, in the joint meta-analysis of the HRS, PGC MDD and the CONVERGE studies, we found that the association between the rs9825823 and the depression continuum (z-score=2.85, P=0.004) was slightly stronger than our initial replication analysis. When these replication and discovery samples were combined, the association with our top hit also became stronger (analyses without the CONVERGE data: z-score= 6.1, P = 1×10−9; with the CONVERGE data z-score= 6.2, P = 6.8×10−10). Results of additional replication analyses are given in the Table S4.
