On the power to detect differences between male and female mutation rates for Duchenne muscular dystrophy, using classical segregation analysis and restriction fragment length polymorphisms.

The power to detect departures from the theoretical proportion of new mutants in X-linked lethal disorders has been analyzed for several types of segregation analysis, including methods based on completely linked restriction fragment length polymorphisms. It is shown that all methods require large sample sizes in order to detect even large differences between male and female mutation rates. Ascertainment bias is shown to have a great effect on the outcome of the segregation analysis. All reviewed studies concerning the proportion of new mutants in Duchenne muscular dystrophy, whether they claimed equality or inequality between the male and female mutation rates, give insufficient evidence because of ascertainment bias and a too low power. An ascertainment bias-free method is given, with the advantage that information from many studies can be combined. By doing so, in the long run, even moderate departures from equality in mutation rates (if present) can be detected.