Comment from Charles Kooperberg and Indika Rajapakse:
'While comparing different approaches to identify interactions in genetic association studies, we obtained significantly different results for the simulations shown in Wu et al. (2010), that would lead to different conclusions about the comparison of their approach to logistic regression. After repeated requests, the authors informed us that they could not provide us with computer code suitable for running simulations. We programmed the method described in this paper. To double check our calculations, both authors of this note programmed the computation of the test statistics independently. We confirmed that our results were similar. The results reported in the attached manuscript are those obtained by the first author - http://www.plosgenetics.o...
