This paper, “Mixed-model association for biobank-scale datasets” (Loh et al. 2018 Nat Genet), explores best practices for genome-wide association analyses in very large cohorts.