Callahan J D, Short J M
Callahan Associates Inc., La Jolla, CA 92037.
Mutat Res. 1995 Mar;327(1-2):201-8. doi: 10.1016/0027-5107(94)00191-7.
Statistical analysis of the lambda/lacI transgenic mutagenicity assay was used to determine optimal sample size and resource allocation in terms of number of animals and number of recovered target genes (recovered phage) required to demonstrate a statistically significant induction in mutant frequency. Statistical assumptions as applied to mutagenicity data are discussed for a number of frequently used statistical analyses. Log transformations are suggested as a means of meeting statistical assumptions and examples are given on interpreting results of analyses of log transformed data. The data analyzed in this study indicate that 300,000 lambda plaques from each of five animals should be analyzed per treatment group in order to detect a doubling of mutant frequencies. Additional sensitivity is gained primarily through increase of animal number and not the number of phage rescued, due to inherent animal-to-animal variability.
采用λ/lacI转基因致突变性分析的统计分析方法，以确定在证明突变频率有统计学显著诱导时，关于动物数量和回收的靶基因（回收噬菌体）数量方面的最佳样本量和资源分配。针对一些常用的统计分析方法，讨论了应用于致突变性数据的统计假设。建议采用对数变换作为满足统计假设的一种方法，并给出了解释对数变换数据的分析结果的示例。本研究分析的数据表明，为了检测到突变频率翻倍，每个处理组应分析来自五只动物的30万个λ噬菌斑。由于动物个体之间存在固有变异性，主要通过增加动物数量而非回收的噬菌体数量来提高额外的灵敏度。
