ABSTRACT

It is no coincidence that the most cited work in the last few years refers to the Comet assay [247]. Yes, the various in vitro and in vivo mutagenicity bioassays make substantial demands on the statistical method. Selected issues are formulated in the following 12 special features:

1. The concentrations used particularly in in vitro assays are somewhat arbitrary in relation to the exposure to humans. Therefore, a tendency to overdosing exists to avoid false negative results. A specific dose-response problem may occur: downturn effects at high(er) concentrations. Trend tests assuming strict order restriction may be seriously biased and therefore a downturn-protected trend test or a one-sided comparison versus control without order restriction (Dunnett-type procedure) is proposed. As an example, Ames assay data are analyzed accordingly; see Section 4.2.