We used available genotyped or imputed SNP information in and around CADM2 (chr 3 (3p12.1), bp 83,951,945‐86,126,470, GRCh37/hg19). Per sample genotyping, imputation and quality control (QC) procedures can be found in Table S1. Variants with a minor allele frequency (MAF) below 1%, a genotype missingness rate above 5%, or deviations from Hardy–Weinberg Equilibrium (HWE) of p < 1e‐10 were excluded from further analysis. SNPs were aligned with the 1,000 Genomes reference panel (phase 3), 37 removing ambiguous SNPs and SNPs that had a MAF that diverged more than 0.15 from that in the reference panel. Following these procedures, n 25Up = 297, n S4S = 2,972, n NTR = 6,166, and n UKB = 4,638 SNPs were available and retained for analysis. Genetic data and data on at least one phenotype were available for N 25Up = 2,133, N S4S = 2,994, N NTR = 12,120, and N UKB = 426,446 individuals (total N = 443,693). The per‐phenotype sample size range was N 25Up = 419–2,071, N S4S = 503–2,384, N NTR = 581–9,432, and N UKB = 23,423–362,018 individuals.
