MEDV描述子预测取代芳烃类化合物的藻毒性

以分子电性距离矢量(MEDV)描述子有效表征43个取代芳烃类化合物的分子结构,应用基于预测变量的选择与模型化(VMSP)方法,建立化合物的封闭体系绿藻(Pseudokirchneriella subcapitata)毒性(48 h EC₅₀)与其分子结构之间的定量相关(QSTR)模型. 应用多元线性回归方法建立的QSTR模型具有较高的相关系数(r0.891 1)及LOO(Leave-One-Out)检验相关系数(q0.810 2),表明该模型具有良好的估计能力与稳定性. 应用28个化合物的训练集样本构建QSTR模型预测外部检验集,结果表明,训练集模型也具有良好的预测能力.      

Cytotoxicity estimation of ionic liquids based on their effective structural features

Cytotoxicity of a diverse set of 227 ionic liquids (taken from UFT/Merck Ionic Liquids Biological Effects Database) containing 94 imidazolium, 53 pyridinium, 23 pyrrolidinium, 22 ammonium, 15 piperidinium, 10 morpholinium, 5 phosphanium, and 5 quinolinium cations in combination with 25 different types of anions to Leukemia Rat Cell Line (IPC-81) was estimated from their structural parameters using quantitative structure - toxicity relationship “QSTR” methodology. Linear and nonlinear models were developed using genetic algorithm (GA), multiple linear regressions (MLR) and multilayer perceptron neural network (MLP NN) approaches. Robustness and reliability of the constructed models were evaluated through internal and external validation methods. Furthermore, chemical applicability domain was determined via leverage approach. In this work, it was revealed that the cationic moieties make the major contribution to cytotoxicity and the anionic parts play a secondary role in cytotoxicity of the ionic liquids studied here. Structural information represented in this work, can be used for a rational design of safer ILs.