Based on effect sizes for the analysis of AO-AD generated in the discovery (COGA) sample, GRS at P-value thresholds of 0.01 (GRS0.01), 0.05 (GRS0.05), 0.10 (GRS0.1) and 0.50 (GRS0.5) were created in SAGE and OZ-ALC sample using PLINK28 and SAS (SAS Institute, Cary, NC, USA). Briefly, SNPs in COGA that were significant at each P-value threshold (for example, P<0.01) were selected. For each SNP, the effect size was calculated as the natural logarithm transformation of the hazard ratio from COGA. For every individual in SAGE and OZ-ALC, this effect size was multiplied by the number of copies of reference allele, and this product was summed across all SNPs.28 For the SAGE dataset, the number of SNPs for each score was: GRSall (110 797), GRS0.5 (58 374), GRS0.1 (12 254), GRS0.05 (6147) and GRS0.01 (1441), while for OZ-ALC number of SNPs for each score was: GPSall (112 594), GPS0.5 (57 053), GPS0.1 (12 161), GPS0.05 (6268) and GPS0.01 (1402). The resulting GRS was used to predict AO-AD, as well as other measures (AO-R, AO-I, AD-SX and Maxdrinks) in those datasets. Associations
