Sahigara Faizan, Ballabio Davide, Todeschini Roberto, Consonni Viviana
Milano Chemometrics and QSAR Research Group, Department of Earth and Environmental Sciences, University of Milano-Bicocca, P.za della Scienza 1 - 20126 Milano, Italy.
Curr Comput Aided Drug Des. 2014;10(2):137-47. doi: 10.2174/1573409910666140410110241.
Several classical and two recently proposed Applicability Domain (AD) approaches were implemented on a set of three classification models retrieved from a published study to assess the ready biodegradability of chemicals. Each model was associated with an optimal AD approach based on its ability to a) retain maximum test molecules within the model's AD, b) be appropriate for the strategy used towards model development and c) show reasonably converging results with those derived with other AD approaches used. A decision criterion was also set to evaluate the AD of two consensus models that were developed in the original study. An overview of test molecules excluded from the AD of all the five biodegradability models was provided including an attempt to identify the major structural features and molecular descriptors possibly relevant in deciding upon their ready biodegradability. Apart from the test set, an overview of the results derived on the external validation set molecules was provided.
在从一项已发表研究中获取的一组三个分类模型上,实施了几种经典的和最近提出的适用域(AD)方法,以评估化学品的易生物降解性。每个模型都与一种最优的AD方法相关联,该方法基于其以下能力:a)在模型的适用域内保留最多的测试分子;b)适用于模型开发所采用的策略;c)与使用其他AD方法得出的结果显示出合理的趋同结果。还设定了一个决策标准来评估在原始研究中开发的两个共识模型的适用域。提供了从所有五个生物降解性模型的适用域中排除的测试分子的概述,包括试图识别在决定其易生物降解性时可能相关的主要结构特征和分子描述符。除了测试集之外,还提供了在外部验证集分子上得出的结果的概述。