A summary of the alcohol-related outcomes included in the analyses is presented in Table 1. Across all samples, participants who did not endorse lifetime alcohol use were removed to account for those with high genetic liability who elect not to drink for personal or socio-cultural reasons, or were not old enough to have had a first drink. Each target sample used a different assessment. Details regarding each assessment are available in the Supplementary Materials. Briefly, COGA participants were administered the Semi-Structured Interview for the Genetics of Alcoholism (SSAGA; (Bucholz et al., 1994)), which includes diagnostic assessments for DSM-IV and DSM-5 AUD as well as other aspects of alcohol use. ALSPAC participants completed the AUDIT questionnaire, and DSM-IV alcohol dependence was derived for 4,328 participants, using symptom-level items adapted from (Kendler, et al., 1992). In GS, participants reported on their alcohol intake (units per week) using a pre-clinical questionnaire, and were also administered the CAGE (Cut-Annoyed-Guilty-Eye) screener (Ewing, 1984). In the UKB, ICD-9 and 10 codes for AUD were derived from linkage to hospital inpatient records. Phenotypic correlations between alcohol use measures across the samples are shown in Supplementary Table 16.
