An evaluation of affected-sib-pair methods and transmission/disequilibrium tests for detecting genes underlying a complex trait

For the analysis of complex traits, it is of interest to compare a few nonparametric methods such as affected-sib-pair (ASP) analyses and transmission/disequilibrium tests (TDT). The affected-sib-pair approaches we have examined here are ASP and ALL-SP which are implemented in SIBPAIR program. We also applied the BETA program which has not so far been extensively compared with other methods. The study indicates that the ASP program and the BETA program give concordant results although BETA tends to give higher lod scores. However, when all sibs were included in the analysis (ALL-SP), linkage signals became weaker, compared with ASP and BETA. The TDT detected 66 positive signals at a significance level of 0.05 and identified a true locus. Overall, our results suggest that affected-sib-pair analysis has reasonable power (p < 0.0001) to detect linkage given the disease model and the family structure specified in the GAW11 Problem 2 data set.