A graph based method for depicting population characteristics using Genome Wide Data

Genome Wide Association (GWA) studies associate genetic variants to clinical phenotypes using statistical tests which are based on assumption of random mating. Populations having history of consanguineous marriages, especially from Middle East, can violate this assumption. Here we present, use of averaged weighted clustering coefficient of undirected graphs to quantify cryptic relatedness between individuals from a random cohort. This measure can be used to understand pattern of relatedness in populations to choose between removing related samples and using mixed linear models correcting for relatedness while performing GWA studies.