Fig. 8
From: A framework for research into continental ancestry groups of the UK Biobank
Graph outlining the possible effects of geographic structure in population genetics: Suppose one might want to use Mendelian randomization to study the relationship between neutrophil count and severe malaria caused by P. Falciparum—a disease largely absent in European environments. Using summary statistics from a neutrophil count GWAS derived from individuals with European ancestry (Fig. 1A) may affect estimates due to geographic structure (Ancestry + Demography + Environment). This can be overcome by running a GWAS in people of African ancestry (Fig. 1B)