Reduction of Sample Heterogeneity through Use of Population Substructure: An Example from a Population of African American Families with Sarcoidosis

摘要: Sarcoidosis is a granulomatous inflammatory disorder of complex etiology with significant linkage to chromosome 5, and marginal was observed five other chromosomes in African Americans (AAs) our previously published genome scan. Because genetic factors underlying disease are often population specific, analysis samples diverse ancestry (i.e., ethnic confounding) can lead loss power. Ethnic confounding addressed by stratifying on self-reported race, controversial less-than-perfect construct. Here, we propose stratified genetically determined as an alternative approach for reducing confounding. Using data from the 380 microsatellite markers genotyped aforementioned scan, clustered AA families into subpopulations basis similarity. Evidence two distinct groups found: subpopulation one (S1) comprised 219 229 families, (S2) consisted six (the remaining four were mixture). Stratified results suggest that only S1 contributed identified signals at 1p22, 3p21-14, 11p15, 17q21 S2 those found 5p15-13 20q13. Signals 2p25, 5q11, 5q35, 9q34 remained both subpopulations, evidence new susceptibility locus 2q37 S2. These demonstrate usefulness ancestry, create homogeneous subsets—more reliable less than race-stratified subsets—in which identify factors. Our findings support presence sarcoidosis-susceptibility genes regions elsewhere but indicate these likely be specific.