Genome-wide association studies (GWAS) have detected hundreds of risk alleles [1], generating novel biological knowledge and widening the range of diagnostic and treatment tools [2]. However, the reported effect sizes of these variants are small and their impact in individual risk prediction remains modest, raising doubts about the relevance of GWAS results [1], [3]–[6]. Some of the most hotly debated topics are how to account for the unexplained risk [4]; what may be the role of rare variants as a source of synthetic GWAS results [7]–[10]; and up to what extent GWAS results are portable between populations [11]–[15].
