Benigni R, Giuliani A
Laboratory of Comparative Toxicology and Ecotoxicology, Istituto Superiore di Sanitá, Rome, Italy.
Environ Health Perspect. 1991 Dec;96:81-4. doi: 10.1289/ehp.919681.
One great obstacle to understanding and using the information contained in the genotoxicity and carcinogenicity databases is the very size of such databases. Their vastness makes them difficult to read; this leads to inadequate exploitation of the information, which becomes costly in terms of time, labor, and money. In its search for adequate approaches to the problem, the scientific community has, curiously, almost entirely neglected an existent series of very powerful methods of data analysis: the multivariate data analysis techniques. These methods were specifically designed for exploring large data sets. This paper presents the multivariate techniques and reports a number of applications to genotoxicity problems. These studies show how biology and mathematical modeling can be combined and how successful this combination is.
理解和利用遗传毒性和致癌性数据库中所包含信息的一个巨大障碍，就是这类数据库的规模本身。其庞大的数据量使其难以解读，进而导致对信息的利用不充分，这在时间、人力和资金方面都代价高昂。奇怪的是，在寻求解决该问题的适当方法时，科学界几乎完全忽略了一系列现有的非常强大的数据分析方法：多元数据分析技术。这些方法是专门为探索大型数据集而设计的。本文介绍了多元技术，并报告了其在遗传毒性问题上的一些应用。这些研究展示了生物学与数学建模如何能够结合，以及这种结合是多么成功。
