Design and analysis of experiments on mutagenicity: I. Minimal sample sizes

Abstract

The probability (power) of collecting sufficient data to demonstrate statistically that a chemical is mutagenic is examined. A formula is derived with which one may calculate the likelihood of correctly concluding a chemical is mutagenic based upon (1) the total sample size, (2) the relative size of the experimental group, (3) the spontaneous mutation rate and (4) the minimal increase in mutation rate that one is concerned in detecting. Figures are also presented which allow one to determine the minimal total sample size required to assure oneself a reasonable chance of successfully demonstrating a chemical to be mutagenic. It is found that the best designed (most powerful) experiments are those with approximately equal numbers of indviduals in the experimental and control groups. Those experiments in which excess numbers of individuals are allotted to either the experimental or control groups suffer a substantial reduction in power. In addition, caution should be exercised when concluding that a chemical is not mutagenic. The inability to collect sufficient data to demonstrate the mutagenicity of a chemical might be attributable to one's using too small a total sample size rather than to lack of mutagenic activity, especially in mutagenicity assay systems involving very small spontaneous mutation rates.