Extended Data Fig. 1: Evaluation of effect estimates and allele frequencies for main GWAS meta-analysis and, in comparison, to the validation analysis.
a,b, Plots showing data for the 299 sentinel variants from the common variant analysis (MAF ≥ 1%) that were also available in the validation set. a, Correlation of allele frequencies between the meta-analysis and the validation cohort from MVP. b, Correlation of effect estimates between meta-analysis and the validation cohort from MVP. The red line is the identity line (x = y). Labelled in red are the variants with discordant direction of effect between meta-analysis and validation. c, Plot showing the relationship between effect allele frequency and strength of effect for sentinel variants of the meta-analysis. The effect estimates are from the inverse variance weighted method for meta-analysis. The dotted vertical lines show the cutoff for rare variants with MAF < 1%. The genome-wide significance cut-off of P < 5 × 10−8 was applied to correct for multiple testing. d, Plot showing co-occurrence of risk allele for atrial fibrillation and minor allele in blue and the inverse in red for sentinel variants of the meta-analysis. AF, allele frequency; ALL, all-ancestry; MAF, minor allele frequency; Meta, meta-analysis; MVP, Million Veteran Program.
