Simulating the Power of Decision Rules Applied to Data From the HGPRT Gene Mutation Assay

A statistician, faced with the problem of evaluating data from mammalian cell gene mutation assays, has to take into account that gene mutations may occur regardless of the activity of the chemical tested. A positive mutant frequency is observed not only in the cell cultures treated by the test compound but also in the controls. Therefore, when the mutant frequency in treatment groups exceeds the spontaneous mutant frequency measured in controls, one has to decide whether this increase is due to random variations or due to a mutagenic effect caused by the test compound. Statistical decision rules may be applied at this stage.