Standard post-GWAS QC was performed on the association results for individual cohorts, the GENOA meta-analysis results, and the gSEM results. For the association results for individual cohorts, the following QC was performed on the summary statistics: (1) variants with MAF < 0.01, imputation quality < 0.8, missing values, or values outside expected ranges were removed; (2) variant IDs were standardized; (3) QQ plots were generated to confirm lack of excessive bias and genomic control was applied to remove residual bias (see Supplementary Table 23 for genomic control factors); and (4) LDSC was performed to confirm correlation among the different OA measures used by cohorts (Supplementary Table 2). For the meta-analysis and gSEM results, the following QC steps were performed: (1) QQ plots were generated to confirm lack of excessive bias and (2) GWS signals were examined for consistency of effect size and direction of effect across cohorts.
