Large, well-phenotyped cohort studies have constituted the backbone of epidemiology for several decades. Prospectively collected longitudinal information on exposures and outcomes enables a broad spectrum of analyses and has led to novel insights into disease etiology, such as the link between smoking and lung cancer [1,2] as well as the link between both high cholesterol levels and trans fatty acids with coronary heart disease [3,4] Many existing cohorts collect biological specimens from their participants, allowing for studies of inherited genetic variation as well as prospectively measured biomarkers such as metabolomic profiles [5] and circulating hormone levels [6]. Genome-wide association studies (GWAS) are currently a main engine of genetic epidemiology and have led to the identification of thousands of loci for hundreds of traits (for an overview and its clinical applications, see Manolio [7]). When designing a GWAS, cost is still the determining factor and consequently, GWAS within cohorts are often conducted within nested case-control studies or sub-cohorts. In contrast, the Women’s Genome Health Study (WGHS) [8] genotyped the entire cohort of 27,000 women and the Genetic Epidemiology Research on Adult
