These data accompany the research detailed in Connally et al. 2021 (https://www.medrxiv.org/content/10.1101/2021.06.08.21258515v2). There are two types of GWAS files in this collection: 1) .linear files are produced by plink - These data were produced using plink2.0 (specifically the --glm command) - Explained at https://www.cog-genomics.org/plink/2.0/formats#glm_linear - Each row is a SNP, and the columns are - CHROM: Chromsome of the variant - POS: Basepair coordinates of the SNP - ID: ID of the SNP - REF: Reference allele - ALT: Alternate allele - A1: Counted allele in the regression - TEST: Test identifier - OBS_CT: Number of samples in the regression - BETA: Regression coefficient - SE: Standare errof of the beta - T_STAT: T-statistic of linear regression - P: P-value of association 2) .cojo files are produced by the COJO method of GCTA64 - Best explained in https://gcta.freeforums.net/thread/178/conditional-joint-analysis-using-summary - Each row is a SNP, and the columns are - Chr: Chromosome of the variant - SNP: ID of the SNP (we used position and alleles as an identifier) - bp: Basepair coordinates of the SNP - refA: Reference allele - freq: Frequency of the effect allele - b: Marginal beta of the allele's effect on the phenotype of choice - se: Standard error of this beta - p: P-value of the SNP's association to phenotype based on the beta and standard error - n: Estimated effective sample size for this SNP - freq_geno: Frequency of the allele in the reference sample - bC: Conditional beta - bC_se: Standard error of the conditional beta - pC: P-value of the conditional effect