Festing M F
Arch Toxicol Suppl. 1986;9:127-37. doi: 10.1007/978-3-642-71248-7_15.
The fundamental principle of the controlled experiment is that treated and control groups should be identical, with minimal within-group variability. Toxicologists recognise this and control age, body weight, disease and the physical environment of the test animals. However most toxicological screening in done with genetically variable outbred stocks, even though isogenic inbred strains, F1 hybrids or identical siblings are usually available. The result is poor experiments with the inevitable genetic differences between groups resulting in increased false positive and negative results, and no indication that the response is under genetic control. It is also illogical to treat genetic variation differently from other types of variation. The argument that it is essential to use outbred animals to model outbred man is illogical. If bacteria can be used to model man (as in the Ames test), so can inbred animals. The uncontrolled variation present in an outbred stock can not be used efficiently to increase the range of phenotypes tested because it also introduces "noise" which obscures experimental effects. The use of two or more isogenic strains gives a much more efficient experimental design with low "noise" and an indication of whether the response is under genetic control. Inbred and F1 hybrid strains (but not identical siblings) have the added advantage of an immortal genotype which outlives any individual animal. Such immortal genotypes may be studied in detail to gather background information. Toxicologists should treat genetics like every other variable and control it, using several isogenic strains in cases where testing needs to be done on more than one genotype.
对照实验的基本原理是，处理组和对照组应完全相同，组内变异性应最小。毒理学家认识到这一点，并控制实验动物的年龄、体重、疾病和物理环境。然而，大多数毒理学筛查是使用基因可变的远交种群进行的，尽管通常可以获得近交系、F1杂种或同卵同胞。结果是实验效果不佳，组间不可避免的基因差异导致假阳性和假阴性结果增加，并且没有迹象表明反应受基因控制。将基因变异与其他类型的变异区别对待也是不合逻辑的。认为必须使用远交动物来模拟远交人群的观点是不合逻辑的。如果细菌可以用来模拟人类（如在艾姆斯试验中），近交动物也可以。远交种群中存在的未控制变异不能有效地用于增加测试表型的范围，因为它还会引入“噪声”，从而掩盖实验效果。使用两个或更多的近交系会得到一个更有效的实验设计，“噪声”低，并且能表明反应是否受基因控制。近交系和F1杂种品系（但非同卵同胞）具有永生基因型的额外优势，这种基因型比任何个体动物的寿命都长。可以对这种永生基因型进行详细研究以收集背景信息。毒理学家应该像对待其他变量一样对待遗传学并加以控制，在需要对不止一种基因型进行测试的情况下，使用多个近交系。
