The approach outlined here could be utilized to expand the case cohort with additional samples of the same disease, even if not directly matched for population, in the same way as controls. However, if gene-environment interaction is likely to exist for the disease under investigation, it is also important to match cases for exposure to potential environmental risk factors, or at least to measure these variables, which may not be possible with publicly available samples. Furthermore, there may be allelic- or genetic-heterogeneity between populations, particularly for complex traits. Under such circumstances, combining case cohorts would reduce power to detect association, and thus would not be advised without prior evidence to the contrary.
