Machine Learning Prediction of Mycobacterial Cell Wall Permeability of Drugs and Drug-like Compounds

Abstract

The cell wall of Mycobacterium tuberculosis and related organisms has a very complex and unusual organization that makes it much less permeable to nutrients and antibiotics, leading to the low activity of many potential antimycobacterial drugs against whole-cell mycobacteria compared to their isolated molecular biotargets. The ability to predict and optimize the cell wall permeability could greatly enhance the development of novel antitubercular agents. Using an extensive structure–permeability dataset for organic compounds derived from published experimental big data (5371 compounds including 2671 penetrating and 2700 non-penetrating compounds), we have created a predictive classification model based on fragmental descriptors and an artificial neural network of a novel architecture that provides better accuracy (cross-validated balanced accuracy 0.768, sensitivity 0.768, specificity 0.769, area under ROC curve 0.911) and applicability domain compared with the previously published results.

Conformational Dynamics and Stability of Bilayers Formed by Mycolic Acids from the Mycobacterium tuberculosis Outer Membrane

“…At the same time, the packing of molecules in a membrane is one of the most important factors that determine the thickness of membranes, their density, the ability to withstand the action of external chemical agents and temperature, as well as the permeability to antibiotics and other small molecules. An alternative approach to the permeability prediction based on the machine learning structure–property models [ 22 ] is also useful but the model applicability domain can be limited by the available data.…”

Free Energy Barriers for Passive Drug Transport through the Mycobacterium tuberculosis Outer Membrane: A Molecular Dynamics Study

“…No such data are available for PRET and MOX molecules; they were considered in this work as test systems for the verification of regression models. CMPI and CMPII are compounds considered by us in our work [25]. They have been shown to have inhibitory effects on Mtb target proteins, but no antibacterial effects on live mycobacteria were found [32][33][34].…”

“…Different machine learning models to predict the Mtb membrane permeability were compared [24]. In the study [25], a predictive in silico model of Mtb cell wall permeability applicable to diverse drugs and drug-like compounds was derived from the extensive Big Data-based dataset. However, it should be borne in mind that the machine learning and regression techniques provide only indirect estimates, and the predicted permeability can be masked by other effects, e.g., the action of the mycobacterial efflux pumps or the action of the enzymes metabolizing the drug molecules inside the living cell.…”