基于地统计学与支持向量回归的QSAR建模

基于主成分分析(PCA)、地统计学(GS)和支持向量回归(SVR), 提出了一种新的定量构效关系(QSAR)个体化预测方法——Weight-PCA-GS-SVR. 其基本思路是: 先以PCA降维并消除自变量间的信息冗余, 继以SVR经非线性主成分筛选去除与因变量无关的主成分, 再以保留主成分计算样本间的加权距离, 然后以高维GS确定公用变程; 每一个待测样本都以自身为中心从训练集中找出加权距离小于公用变程的私有k个近邻, 以SVR训练建模完成个体化预测. Weight-PCA-GS-SVR从行、列两个方向对模型进行了优化, 为自变量提供了一种新的加权方法, 为解决最优k近邻选择难题提供了新的思路, 并具有SVR原来的优点. 经3个化合物活性实例数据集验证, 新方法在所有参比模型中预测精度最高, 且明显优于文献报道结果, Weight-PCA-GS-SVR在QSAR等回归预测领域有较广泛的应用前景.     

Three useful dimensions for domain applicability in QSAR models using random forest

One popular metric for estimating the accuracy of prospective quantitative structure-activity relationship (QSAR) predictions is based on the similarity of the compound being predicted to compounds in the training set from which the QSAR model was built. More recent work in the field has indicated that other parameters might be equally or more important than similarity. Here we make use of two additional parameters: the variation of prediction among random forest trees (less variation among trees indicates more accurate prediction) and the prediction itself (certain ranges of activity are intrinsically easier to predict than others). The accuracy of prediction for a QSAR model, as measured by the root-mean-square error, can be estimated by cross-validation on the training set at the time of model-building and stored as a three-dimensional array of bins. This is an obvious extension of the one-dimensional array of bins we previously proposed for similarity to the training set [Sheridan et al. J. Chem. Inf. Comput. Sci.2004, 44, 1912-1928]. We show that using these three parameters simultaneously adds much more discrimination in prediction accuracy than any single parameter. This approach can be applied to any QSAR method that produces an ensemble of models. We also show that the root-mean-square errors produced by cross-validation are predictive of root-mean-square errors of compounds tested after the model was built.     

Selection of appropriate training and validation set chemicals for modelling dermal permeability by U-optimal design

Quantitative structure-activity relationship (QSAR) models are being used increasingly in skin permeation studies. The main idea of QSAR modelling is to quantify the relationship between biological activities and chemical properties, and thus to predict the activity of chemical solutes. As a key step, the selection of a representative and structurally diverse training set is critical to the prediction power of a QSAR model. Early QSAR models selected training sets in a subjective way and solutes in the training set were relatively homogenous. More recently, statistical methods such as D-optimal design or space-filling design have been applied but such methods are not always ideal. This paper describes a comprehensive procedure to select training sets from a large candidate set of 4534 solutes. A newly proposed ‘Baynes’ rule’, which is a modification of Lipinski's ‘rule of five’, was used to screen out solutes that were not qualified for the study. U-optimality was used as the selection criterion. A principal component analysis showed that the selected training set was representative of the chemical space. Gas chromatograph amenability was verified. A model built using the training set was shown to have greater predictive power than a model built using a previous dataset [1].     

Constructing optimum blood brain barrier QSAR models using a combination of 4D-molecular similarity measures and cluster analysis

A new method, using a combination of 4D-molecular similarity measures and cluster analysis to construct optimum QSAR models, is applied to a data set of 150 chemically diverse compounds to build optimum blood-brain barrier (BBB) penetration models. The complete data set is divided into subsets based on 4D-molecular similarity measures using cluster analysis. The compounds in each cluster subset are further divided into a training set and a test set. Predictive QASAR models are constructed for each cluster subset using the corresponding training sets. These QSAR models best predict test set compounds which are assigned to the same cluster subset, based on the 4D-molecular similarity measures, from which the models are derived. The results suggest that the specific properties governing blood-brain barrier permeability may vary across chemically diverse compounds. Partitioning compounds into chemically similar classes is essential to constructing predictive blood-brain barrier penetration models embedding the corresponding key physiochemical properties of a given chemical class.     

Local and global quantitative structure-activity relationship modeling and prediction for the baseline toxicity

The predictive accuracy of the model is of the most concern for computational chemists in quantitative structure-activity relationship (QSAR) investigations. It is hypothesized that the model based on analogical chemicals will exhibit better predictive performance than that derived from diverse compounds. This paper develops a novel scheme called "clustering first, and then modeling" to build local QSAR models for the subsets resulted from clustering of the training set according to structural similarity. For validation and prediction, the validation set and test set were first classified into the corresponding subsets just as those of the training set, and then the prediction was performed by the relevant local model for each subset. This approach was validated on two independent data sets by local modeling and prediction of the baseline toxicity for the fathead minnow. In this process, hierarchical clustering was employed for cluster analysis, k-nearest neighbor for classification, and partial least squares for the model generation. The statistical results indicated that the predictive performances of the local models based on the subsets were much superior to those of the global model based on the whole training set, which was consistent with the hypothesis. This approach proposed here is promising for extension to QSAR modeling for various physicochemical properties, biological activities, and toxicities.     

Kinase-kernel models: accurate in silico screening of 4 million compounds across the entire human kinome

Reliable in silico prediction methods promise many advantages over experimental high-throughput screening (HTS): vastly lower time and cost, affinity magnitude estimates, no requirement for a physical sample, and a knowledge-driven exploration of chemical space. For the specific case of kinases, given several hundred experimental IC(50) training measurements, the empirically parametrized profile-quantitative structure-activity relationship (profile-QSAR) and surrogate AutoShim methods developed at Novartis can predict IC(50) with a reliability approaching experimental HTS. However, in the absence of training data, prediction is much harder. The most common a priori prediction method is docking, which suffers from many limitations: It requires a protein structure, is slow, and cannot predict affinity. (1) Highly accurate profile-QSAR (2) models have now been built for roughly 100 kinases covering most of the kinome. Analyzing correlations among neighboring kinases shows that near neighbors share a high degree of SAR similarity. The novel chemogenomic kinase-kernel method reported here predicts activity for new kinases as a weighted average of predicted activities from profile-QSAR models for nearby neighbor kinases. Three different factors for weighting the neighbors were evaluated: binding site sequence identity to the kinase neighbors, similarity of the training set for each neighbor model to the compound being predicted, and accuracy of each neighbor model. Binding site sequence identity was by far most important, followed by chemical similarity. Model quality had almost no relevance. The median R(2) = 0.55 for kinase-kernel interpolations on 25% of the data of each set held out from method optimization for 51 kinase assays, approached the accuracy of median R(2) = 0.61 for the trained profile-QSAR predictions on the same held out 25% data of each set, far faster and far more accurate than docking. Validation on the full data sets from 18 additional kinase assays not part of method optimization studies also showed strong performance with median R(2) = 0.48. Genetic algorithm optimization of the binding site residues used to compute binding site sequence identity identified 16 privileged residues from a larger set of 46. These 16 are consistent with the kinase selectivity literature and structural biology, further supporting the scientific validity of the approach. A priori kinase-kernel predictions for 4 million compounds were interpolated from 51 existing profile-QSAR models for the remaining >400 novel kinases, totaling 2 billion activity predictions covering the entire kinome. The method has been successfully applied in two therapeutic projects to generate predictions and select compounds for activity testing.     

基于机器学习方法的丙型肝炎病毒聚合酶NS5B非核苷抑制剂的定量构效关系研究

对89 个苯并异噻唑和苯并噻嗪类丙型肝炎病毒(HCV) NS5B聚合酶非核苷抑制剂进行了定量构效关系(QSAR)研究. 采用遗传算法组合偏最小二乘(GA-PLS)和线性逐步回归分析(LSRA)两种特征选择方法选择最优描述符子集, 然后建立多元线性回归和偏最小二乘线性回归模型. 并首次尝试使用遗传算法耦合支持向量机方法(GA-SVM)对两种特征选择方法所选的描述符子集分别建立非线性支持向量机回归模型. 三种机器学习方法所建模型均得到比较满意的预测效果. 采用LSRA所选的6 个描述符建立的三个QSAR模型对于测试集的相关系数为0.958-0.962, GA-SVM法给出最好的预测精度(0.962). 采用GA-PLS所选的7个描述符建立的三个QSAR模型对于测试集的相关系数为0.918-0.960, 偏最小二乘回归模型的结果最好(0.960). 本工作提供了一种有效的方法来预测丙型肝炎病毒抑制剂的生物活性, 该方法也可以扩展到其他类似的定量构效关系研究领域.