|
Abstract:
|
Both family- and population-based designs can provide powerful inference for studying associations between genetic variants and risk, but both have major limitations. A hybrid design, best for rare diseases with onset early in life, augments a set of cases and their parents with population-based controls and their parents. The cases and their parents are genotyped, but only parents of controls are genotyped, while exposures are ascertained for both. A log-linear analysis allows estimation of relative risk parameters, provides more power than either the case-parents approach or a case-mother/control-mother approach, allows for missing genotypes, permits direct testing for bias due to mating asymmetry or population stratification, and permits valid alternative analyses when such bias is in evidence. Gene-by-environment interaction also can be assessed more efficiently using this design.
|
