My Submission Package

Abstract

Abstract Both polygenicity 1,2 ( i.e. many small genetic effects) and confounding biases, such as cryptic relatedness 3 population stratification polygenicity , can yield inflated distributions of test statistics in genome-wide association studies (GWAS). However, current methods cannot distinguish between inflation from bias many true signal and polygenicity. We have developed an approach that quantifies the contributions and each by examining and relationship from and confounding the linkage disequilibrium (LD). can term this of LD Score regression. of of regression provides Score upper bound on LD contribution of test test to test observed statistics statistics in in between inflation be used inflation estimate a more powerful correction factor than genomic control 4–14 . approach find strong evidence bias to accounts for We majority We an statistic that the the GWAS the large sample size.