From the onset of the pandemic caused by the virus SARS-CoV-2, the scientific community responded– with a sense of urgency – by intensifying efforts to provide drugs effective against the disease COVID-19. To strengthen this efforts, a consortium of researchers initiated in March 2020 the “COVID Moonshot project” that has been accepting public suggestions for computationally triaged, synthesized, and tested molecules, with experimental data made publicly available. The main goal of the project was to identify through Fragment Based Drug Design (FBDD) small molecules with activity against the virus, for oral treatment. Since orally administered drugs are introduced to the bloodstream through absorption via the small intestine pathway, the ability of a drug to readily cross the intestinal cell membranes and enter circulation is decisively influencing its bioavailability. This explains the need to evaluate and optimize a drug’s membrane permeability in the early stages of drug discovery to avoid failures in late-stage drug development owing to incomplete absorption and poor bioavailability. In our present work, as a contribution to the ongoing scientific efforts, we have employed advanced Machine Learning techniques, including stacked model ensembles, to develop QSAR tools for modelling the PAMPA Effective Permeability (passive diffusion) of orally administered drugs. By applying feature elimination methods, we identified a set of 61 features (descriptors) most relevant in explaining drug cell permeability and used these features to develop the models. The QSAR models were subsequently used to predict the PAMPA Effective Permeability of molecules included in datasets made available through the COVID Moonshot project. Our models were shown to be robust and may provide a promising framework for predicting the potential permeability of molecules not yet synthesized, thus guiding the process of drug design.
Keywords: covid-19, PAMPA, permeability, QSAR, ensemble modelling, descriptors
|Format: XLSX||Size: 1.2 MB||Download|
When a peer-reviewed version of this preprint is available, this information will be updated in the information box above. If no peer-reviewed version is available, please cite this preprint using the following information:
Gousiadou, C.; Doganis, P.; Sarimveis, H. Beilstein Arch. 2021, 202122. doi:10.3762/bxiv.2021.22.v1
|Download RIS (Reference Manager)||Download BIB (BIBTEX)|