Orozco et al. [16], reviewing empirical and theoretical data, and drawing particularly upon evidence from linkage studies, pathway analyses and trans-ethnic studies, also recently concluded that synthetic associations with multiple rare variants cannot be responsible for many reported GWAS associations. The synthetic association hypothesis was also specifically addressed in a recent multi-ethnic study. Waters et al. [17] took 19 variants (15 had RAF>0.3) reproducibly associated to Type 2 Diabetes in Europeans and tested them in 5 racial/ethnic groups (European Americans, African Americans, Latinos, Japanese Americans, and Native Americans). Despite the relatively small sample size for what was essentially a replication study (total size of 14,000 across all groups) the OR for total number of risk alleles in each ethnic group was highly significant, implying ancient causal variants predating the migrations that separated these populations. Waters et al. argued that synthetic associations with rare variants could not explain these common associations with Type 2 diabetes.
