Large-sample pairwise comparisons among multinomial proportions with an application to analysis of mutant spectra

Abstract

In studies examining patterns of mutational damage, the primary variables of interest are expressed typically as discrete counts within defined categories of damage. The multinomial distribution is a common model for such data settings. Of interest is statistical comparison of the pattern, or spectrum, of mutation among the various mutant categories. A specific question in such instances is the identification of differences among the spectral categories, identified, e.g., via simultaneous confidence intervals for pairwise differences in or ratios of the multinomial response probabilities. Herein, a selection of possible methods for constructing such intervals is described and compared via computer simulations to determine which are most appropriate for practical use in the evaluation of mutant spectra.