In advancing epidemiological study, the improvement of both cohort study and biobank is essential. Also, the development of statistical genetics approaches to identify disease risk factors is indispensable. In this article, we will review the recent biobank and cohort study to promote epidemiological research, and introduce the Tohoku Medical Megabank Project, which includes the largest three‐generation birth cohort in the world. Next, we describe the “missing heritability”, which is one of the most difficult issue in cohort research, and show one of the hypotheses, the polygenic model, to overcome this. In addition, we introduce the iWate PolyGenic Model (iPGM) based on the polygenic model to predict genetic predisposition to Ischemic stroke. Finally, we discuss effective approaches toward prediction of multifunctional disorders such as psychiatric disorders through cohort study.