A database of results has been created using chemicals tested by NTP for mutagenesis in short-term assays, and for carcinogenesis. This database allows the evaluation of each short-term assay with respect to its ability to predict carcinogenesis or other short-term assay results. It also permits studies of the individual assays with respect to inter- and Intra-laboratory reproducibility. A study to examine the predictivity of the four short-term tests as a function of the potency of the short-term test and carcinogenicity test responses is continuing. The chemicals to be evaluated include the 114-chemical dataset used , previously to correlate qualitative short-term test results with carcinogenicity results, which is being used to correlate the same parameters using the potencies of the test responses. These measures of potency are being developed and evaluated, and will be used for the correlational studies. The predictivities for Salmonella mutagenicity of two computer-based structure-activity systems, one empirical structure-activity system, and one physicochemical system, were determined. The test database contained 100 chemicals selected at random from among chemicals that had teen tested for mutagenicity in Salmonella, but whose results had not been published by the NTP. The chemical structures were drawn and sent, without identifying them by name, or coded chemicals, to the participants. The purpose of the study was threefold: to measure the accuracy of the different test systems as they were configured at the time of the study: to use the results of this exercise to identify the strengths and weaknesses' of each system: and to bring understanding and light to structure-activity predictions. The two computer-based SAR systems and the empirical system gave equivalent predictivities (approx. 70-80%), while the physicochemical system was not as effective (approx. 60% predictivity). The analyses of the results, and a determination of the differences and similarities among the various systems is continuing.
