An often challenging and sometimes underappreciated facet of genetic epidemiology is the process of choosing an appropriate definition of the trait of interest. Consider for example Alzheimer's disease. While advanced age is the main risk factor for this disease, and it most often strikes after age 65, there are people that develop this disease before the age of 50. Pathologically, the disease appears the same regardless of the age of onset: microscopic examination of brain tissue from patients that are struck with this disease at any age reveal the presence of both amyloid plaques and neurofibrillary tangles. However, it is now recognized that the patients with the very early ages of onset have a mutation in one of three genes (APP, PS1, PS2), and patients that develop Alzheimer's disease at a later age lack these mutations [1,2]. Early-onset Alzheimer's disease is inherited as a Mendelian dominant trait, while the more common form of the disease is a complex trait with a significant environmental component. Thus, any genetic analysis of Alzheimer's disease that did not include age of onset as a
