Today’s scenario is more of a Chemo-centric led ligand based approach aided by Artificial Intelligence (AI), as it has been gaining importance steadily in the drug discovery process. This approach is based on the hypothesis of using the chemical similarity among ligand sets as a representation for the pharmacological similarities of the protein targets and hence would quantify target similarity on the basis of similarities of ligands.
There has been a lot of curated data pointing towards the implementation of chemo-centric approach aided by AI methods. For instance when available targets with highly interrelated sequences were found to be pointing towards dissimilar ligands and dissimilar receptors by sequence are pointing towards similar ligands; enhances the idea of poly pharmacological concepts that could be incorporated in the AI techniques for virtual screening by using ligand based approach in order to ensure better screening . Thus, selection of key descriptors is an important step in any of the AI methods being used. Finding and identification of patterns (predictive fingerprints or combinations of features) that correlate with activity is essential. Furthermore, compounds exhibiting promising properties may be compared with other candidates in order to identify other potential drugs that share critical features.
Therefore, it is evident that the AI approaches to feature-selection, pattern-recognition, classification and clustering could be applied to the various drug discovery related screening for Mtb inhibitors. It has also been reported by OSDD consortium that Random Forest performs well in the classification of imbalanced data set with less diverse samples. But Ekins S etal., preferred Bayesian inference networks for the implementation of similarity-based virtual screening, as it has been found to provide an effective tool for ligand-based virtual screening . Therefore repositioning of chemical compounds from different classes provides some hope to hits recognition as a part of virtual screening.