Physicochemical property profile for brain permeability: comparative study by different approaches

Prediction of the Blood-Brain Barrier (BBB) Permeability of Chemicals Based on Machine-Learning and Ensemble Methods

The ability of chemicals to enter the blood-brain barrier (BBB) is a key factor for central nervous system (CNS) drug development. Although many models for BBB permeability prediction have been developed, they have insufficient accuracy (ACC) and sensitivity (SEN). To improve performance, ensemble models were built to predict the BBB permeability of compounds. In this study, in silico ensemble-learning models were developed using 3 machine-learning algorithms and 9 molecular fingerprints from 1757 chemicals (integrated from 2 published data sets) to predict BBB permeability. The best prediction performance of the base classifier models was achieved by a prediction model based on an random forest (RF) and a MACCS molecular fingerprint with an ACC of 0.910, an area under the receiver-operating characteristic (ROC) curve (AUC) of 0.957, a SEN of 0.927, and a specificity of 0.867 in 5-fold cross-validation. The prediction performance of the ensemble models is better than that of most of the base classifiers. The final ensemble model has also demonstrated good accuracy for an external validation and can be used for the early screening of CNS drugs.

Contribution assessment of multiparameter optimization descriptors in CNS penetration

Assessment of the influence of six physicochemical properties used in the multiparameter optimization (MPO) approach for chemical penetration of the blood-brain barrier was carried out by means of application of logistic regression and multiple linear regression, using a data set of 578 diverse chemicals. It was found that use of an aggregation MPO-score descriptor did not give satisfactory results with central nervous system (CNS)/non-CNS classification. Thus an application of the MPO approach for CNS penetration is ambiguous. An alternative to the MPO approach in this work contains detailed (quantitative) structure-activity relationship analysis using a number of methods (linear discriminant analysis, random forest, support vector machine, Gaussian process). Three properties (molecular weight, number of H-bond donors and octanol-water partition coefficient) yielded optimal categorical models with modest statistical parameters (accuracy 0.730-0.765 for CNS/non-CNS classification). The poor statistics of regression models for the common data set suggested the presence of subsets with different mechanisms of penetrations. Based on graphic comparison of experimental and calculated C_u,b values, subset clusters have satisfactory statistics. The regression models obtained allowed the estimation of descriptor contributions in log C_u,b. This means that medicinal chemists now have a simple additive scheme for at least preliminary quantitative assessment of this important pharmacokinetic parameter.

Assessment of the classification abilities of the CNS multi-parametric optimization approach by the method of logistic regression

Assessment of "CNS drugs/CNS candidates" classification abilities of the multi-parametric optimization (CNS MPO) approach was performed by logistic regression. It was found that the five out of the six separately used physical-chemical properties (topological polar surface area, number of hydrogen-bonded donor atoms, basicity, lipophilicity of compound in neutral form and at pH = 7.4) provided accuracy of recognition below 60%. Only the descriptor of molecular weight (MW) could correctly classify two-thirds of the studied compounds. Aggregation of all six properties in the MPOscore did not improve the classification, which was worse than the classification using only MW. The results of our study demonstrate the imperfection of the CNS MPO approach; in its current form it is not very useful for computer design of new, effective CNS drugs.