Abstract
Cryptic relatedness is inherently a feature of large genome-wide association studies (GWAS), and can give rise to considerable inflation in summary statistics through variance distortion. It has proven difficult to disentangle these inflationary effects from true polygenic effects. Uncorrected inflation will alter the estimates of the number of causal loci associated with the phenotype and the strength of their association. Here we present results of a model that enables estimation of polygenicity, strength of association (or discoverability), and residual inflation in GWAS summary statistics. We show that there is residual inflation (“true” genomic control) of and for recent large GWAS of height and schizophrenia, respectively. In contrast, a larger GWAS for educational attainment shows no residual inflation. Additionally, we find that height has a relatively low polygenicity, with approximately 8k SNPs having causal association, more than an order of magnitude less than has been reported. Residual inflation in GWAS summary statistics can be detected and then corrected using the genomic control procedure with in place of the traditional genomic control factor.