There is scientific and regulatory interest in using mutation assays in transgenic mice in safety assessments for new chemicals and drugs. Currently these assays are in the process of being validated, and protocols for routine testing are being defined. Some of the issues and results to date with regard to assay validation include reproducibility of the assay results (they are qualitatively reproducible), relevance of the test system (the transgene closely approximates an endogenous mammalian gene as a mutational target for the limited number of compounds tested), and the predictivity of the assay for heritable effects (unknown at this time) or carcinogenicity (the assays show good positive predictivity for carcinogenicity; the negative predictivity of the assay requires further investigation). Definition of appropriate study protocols for routine testing requires that applicable statistical methods are available and that the experimental parameters that affect the detection of mutations are known. Progress made in identifying these parameters is discussed. A proposal is made for the custom design of routine safety studies, which is based on the anticipated use of each individual test agent. A working group has been formed to conduct some of the studies still required for validation of these assays.