FinnGen Database

The publicly available FinnGen database contains genotype data from Finnish biobanks linked with corresponding national health registry data from 12,859 AF cases and 73,341 controls. Data from the FinnGen database was retrieved from freeze number 3. FinnGen participants were genotyped by the FinnGen investigators using various Illumina and Affymetrix GWAS arrays. This was followed by a genotype imputation by the Beagle v.4.1 software using 3,775 high coverage whole genome sequences of Finnish origin as a population specific reference panel (6). We used the FinnGen disease endpoint “Atrial Fibrillation and Flutter.” The SAIGE software was used for the genetic association tests of the FinnGen data. SAIGE is a mixed model logistic regression R/C++ package that is able to adjust for relatedness of the analyzed samples while also controlling for unbalanced case-control ratios that can otherwise lead to inflated type 1 error rates (7). SAIGE first generates relatedness estimates and a null model. Next, the null model is fitted with the respective SNP, and the genetic kinship matrix and the model itself is then adjusted for the following covariates: sex, age, 10 principal components and genotyping batch. For more information on the included biobanks, disease endpoint definitions and genotype data in the FinnGen database, please visit https://finngen.gitbook.io/documentation/v/r3/.