In Silico Prediction of Human Intestinal Permeability (Caco-2) using QSPR Modelling for Efficient Drug Discovery.

Background: The quantitative structure-property relationship (QSPR) modelling can be helpful in the in-silico prediction and pre-determination of the drug permeability values of a large number of compounds through human intestinal enterocytes for screening of potential candidate drugs, thereby enhancing oral drug development.

Methods: The present study involved the development of a regression-based QSPR model for the prediction of Caco-2 cell-permeability values of compounds. The training of the model was carried out on a novel large dataset of 1272 compounds with 30 selected 2D descriptors.

Results: An R2 value of 0.96 suggested that the model was significant. Finally, the model was applied in the virtual screening of 49,430 potential compounds of the CAS database of antiviral compounds, among which the model successfully screened 100 compounds as potential leads, with 96 compounds falling within the Applicability Domain (AD).

Conclusion: The present study highlights in-silico screening, which could be beneficial for the early stages of drug development.