Effect of mode of inheritance when calculating the power of a transmission/disequilibrium test study

In assessing the power of an association study one needs to consider a variety of genetic models and to have some knowledge of the values of the relevant population parameters in those models. One's comparison of different models should be made such that these parameters are equal under each model. In practice one is most likely to have available estimates of the allele frequency (p) and of the relative risk to siblings of affected individuals (lambda(s)). In presenting power calculations for the transmission/disequilibrium test, previous authors have made comparisons of models which, while having the same gene frequency lead to completely different population outcomes, e.g. different values of lambda(s). Indeed, despite having the same values for their underlying parameters, these models do not necessarily provide the most useful comparison.