Benigni R
Istituto Superiore di Sanità, Rome, Italy.
Mutagenesis. 1986 May;1(3):185-90. doi: 10.1093/mutage/1.3.185.
Computerized data analysis methods were applied to a selected subset of the results generated by the International Program on Chemical Safety Collaborative Study on In Vitro Assays. The rationale of this study was the analysis of test performances and the comparison with the conclusions of a similar analysis previously performed on the results of the First International Study. The multi-variate data analysis techniques used in the present work proved to be particularly suitable for looking at the complex field of comparative mutagenesis in a logical and systematic fashion, and generated a clear classification of the assays on the basis of their experimental performances. This classification proved not to be consistent with the traditional categories that group tests according to genetic end-point and phylogenetic position, and afforded classes of assays distinct for their sensitivity and specificity in respect to carcinogens. In particular, two systems (Saccharomyces cerevisiae XV185-14C mutation and Syrian Hamster Embryo cells transformation assays) on the basis of the published results, seemed to show a considerable sensitivity in predicting carcinogenicity, although they do not show high specificity.
计算机数据分析方法应用于国际化学品安全规划署体外试验协作研究产生的选定结果子集。本研究的基本原理是分析测试性能，并与先前对第一次国际研究结果进行的类似分析结论进行比较。本研究中使用的多变量数据分析技术被证明特别适合以逻辑和系统的方式审视比较诱变这一复杂领域，并根据实验性能对试验进行了清晰分类。事实证明，这种分类与根据遗传终点和系统发育位置对试验进行分组的传统类别不一致，并且根据对致癌物的敏感性和特异性提供了不同的试验类别。特别是，根据已发表的结果，两个系统（酿酒酵母XV185 - 14C突变试验和叙利亚仓鼠胚胎细胞转化试验）在预测致癌性方面似乎表现出相当高的敏感性，尽管它们的特异性不高。
