Lipnick R L
US Environmental Protection Agency, Office of Toxic Substances (TS-796), Washington, DC 20460.
Sci Total Environ. 1991 Dec;109-110:131-53. doi: 10.1016/0048-9697(91)90175-e.
Quantitative structure-activity relationships (QSARs) provide a useful tool for defining a mathematical relationship between chemical structure and toxicity, and for applying such statistically derived models for predicting the toxicity of untested chemicals. Outlier chemicals can be encountered both in the derivation of QSAR models as well as in their application. Information regarding the relationship between molecular descriptors and molecular mechanism of toxicity can provide insight into the origin of outlier behavior as well as guidance regarding the predictive limitations of such models. Comparison of measured toxicity data for fish and rats with baseline QSAR prediction provides a means of identifying outliers and for categorizing more specific molecular mechanisms.
定量构效关系（QSARs）为定义化学结构与毒性之间的数学关系，以及应用此类统计推导模型预测未测试化学品的毒性提供了有用的工具。在QSAR模型的推导及其应用过程中都可能遇到异常化学品。有关分子描述符与毒性分子机制之间关系的信息可以深入了解异常行为的起源，以及此类模型预测局限性的相关指导。将鱼类和大鼠的实测毒性数据与基线QSAR预测进行比较，提供了一种识别异常值并对更具体的分子机制进行分类的方法。
