Multilocus linkage analysis of the German asthma data

We analyzed data from the German Asthma Genetics Group with three methods t hat utilize pedigree-specific nonparametric linkage scores to facilitate th e search for multiple independent and interacting susceptibility loci. The three methods included a conditional analysis, logistic regression, and neu ral networks. Although there were differences, the three methods identified many of the same susceptibility loci. The most consistent evidence was pro vided for loci on chromosomes 1, 2, 6, 9, and 15. Both the conditional and the logistic regression analyses suggested an epistatic relationship betwee n loci on chromosomes 2 and 9. The logistic regression analysis further rev ealed evidence for locus heterogeneity between loci on chromosomes 6 and 15 . Finally, the neural network analysis identified a potential locus on chro mosome 17 that was not identified in the other analyses. (C) 2001 Wiley-Lis s, Inc.