View Program - 2022 Population, Evolutionary, and Quantitative Genetics Conference

Population stratification is a well-studied problem in genome-wide association studies, leading to biases in the estimated strength of phenotypic association for individual genetic variants. In short, if environmental effects on the phenotype are correlated with ancestry gradients within a GWAS panel, any variant that is stratified along this ancestry gradient will receive a biased effect size estimate. While state of the art methods to correct for stratification are generally effective in reducing the number of significant false positive associations, even subtle biases in effect size estimates can accumulate across loci, leading to systematic biases in polygenic scores. In turn, these biases in the distribution of polygenic scores can lead to false positives in downstream analyses, such as tests for polygenic adaptation or other analyses of among group genetic differences. One approach is to be extremely aggressive in the use of fixed effects genetic PCs to control for stratification. However, there is currently no way to tell conclusively if confounding effects have been removed. A second approach is to conduct the GWAS in an evolutionarily diverged sample that is less likely to share population genetic structure with the test panel. This renders potential biases in the effect sizes irrelevant to the test, but comes at the cost of significantly reduced statistical power due to the poor portability of polygenic scores across samples of divergent ancestry. Using theory from population and statistical genetics, together with simulations, we show that even if GWAS and test panels do share genetic structure it is possible to guarantee the unbiasedness of polygenic selection tests without needing to achieve the much more difficult task of guaranteeing that the effect sizes are completely unbiased. We show that by analyzing GWAS and test panels jointly in a unified framework, we can leverage the observed overlap in population structure between the two samples so as to protect the GWAS from stratification biases along the relevant axis of shared structure. More generally, our results have implications beyond tests for selection as any analysis that attempts to quantify the covariance between polygenic scores and demographic or environmental variables is subject to the same type of stratification biases, and can therefore benefit from our framework.

Guaranteeing unbiasedness in selection tests based on polygenic scores