【摘要】 本文把序贯判别方法应用于以遗传毒理学试验预测化学品的致癌性,优化预测的社会效益。给出了遗传毒理学试验组合的较佳选择方案以及实用的序贯判别方法。推荐以沙门氏菌回变试验为第一个试验,啮齿类骨髓微核试验为第二个试验,哺乳动物细胞SCE试验或哺乳动物细胞UDS试验或啤酒酵母有丝分裂重组试验为第三个试验。本文还改进了Titterington处理缺失数据的统计方法。更多还原

【Abstract】 The discriminant tree method has been used in the prediction of carcinogenicity of chemicals withgenotoxicity assays so that the best social effectiveness could be achieved. The optimal batteries areselected based on the genotoxicity data base and a feasible sequential discriminant method is given.Results of the analysis suggest that the optimal or nearly optimal batteries consist of Salmonella reversemutation test (the 1st assay), rodent micronucleus test (the 2nd assay) and Saccharomyces cerevisiaemitotic recombination test or mammalian cells SCE test in vitro or mammalian cells UDS test in vitro(the 3rd assay). Titterington’s method dealing with incomplete data is improved and given in theAppendix.更多还原