Combining Machine Learning Systems and Multiple Docking Simulation Packages to Improve Docking Prediction Reliability for Network Pharmacology

Increased availability of bioinformatics resources is creating opportunities for the application of network pharmacology to predict drug effects and toxicity resulting from multi-target interactions. Here we present a high-precision computational prediction approach that combines two elaborately built machine learning systems and multiple molecular docking tools to assess binding potentials of a test compound against proteins involved in a complex molecular network. One of the two machine learning systems is a re-scoring function to evaluate binding modes generated by docking tools. The second is a binding mode selection function to identify the most predictive binding mode. Results from a series of benchmark validations and a case study show that this approach surpasses the prediction reliability of other techniques and that it also identifies either primary or off-targets of kinase inhibitors. Integrating this approach with molecular network maps makes it possible to address drug safety issues by comprehensively investigating network-dependent effects of a drug or drug candidate.

Figure: Illustration of a systems docking simulation. It is mainly used to dock a small molecule to a protein structure and to evaluate its potential to bind.

Docking simulation, a screening method to rapidly assess a test compound's binding activity, is especially helpful in early stage pharmacology studies. We developed a docking method using machine learning approach, which integrates features of structure-based rational drug design and QSAR into the learning models to enhance performance in molecular recognition. Docking simulation conducted by machine learning systems A + B provides improved reliability in predicting binding potentials and the capability of identifying potential targets. To achieve more accurate prediction, further integration of other computer-aided technology is feasible, such as the application of molecular dynamics (MD) after docking. Together with a curated signaling map, the network-based screening approach is able to comprehensively characterize the underlying mechanism of a drug candidate's activity and also to interpret the cascade effects of modulated targets. Adverse side effects constitute an enormous cost in drug development. By applying network-based screening, drug developers can reduce the possibility of marketing a drug with unfavorable drug-target interactions. On the other hand, it also provides an opportunity to rationally optimize inhibitor polypharmacology for treating complex diseases, such as cancer, neurodegenerative disorders, cardiovascular disease, and metabolic syndromes.