Non-insulin-dependent diabetes mellitus, or type 2 diabetes (T2DM), has become a major public health problem in India. The high prevalence, relatively young age of onset, and strong familial aggregation of T2DM in some Indian communities remains to be explained. Many of the traditional risk factors established for European populations do not appear to be present in Asian Indians. Phase I of the Sikh Diabetes Study (SDS) was launched to build the population resources required to initiate a large-scale genetic epidemiological study of diabetes in an Asian Indian population. The SDS is focused on the Khatri Sikh population of North India. In all, 1,892 subjects were enrolled to participate in the family-based study; 1,623 of these subjects belong to 324 families, each of which has at least 2 siblings affected with T2DM. The sample included 1,288 individuals affected with T2DM (siblings, parents, or relatives) and 335 unaffected siblings, parents, or relatives. The remaining 269 subjects were unrelated nondiabetic control subjects, including unaffected spouses of probands or siblings. This primarily nonvegetarian, nonsmoking endogamous caste group has presented an unusual clinical picture of uneven distribution of adiposity and a high rate of T2DM with secondary complications. Such well-characterized population isolates may offer unique advantages in mapping genes for common complex diseases.