Probing helminth proteome for anti-hypertensive small natural peptides (small peptide of <2000 kDa) inhibitor for ACE2/ Cathepsin L to inhibit SARS-CoV-2 entry
The first step is Data mining to generate a de novo peptide library from helminth and creating a training data set with existing anti-hypertensive peptide drugs. The literature for anti-hypertensive drugs is searched and the database is curated for a training set and evaluated using BIOPEP. The helminth proteins are downloaded from the database (worm.org) and protein set is used to generate the de novo peptide library. Performing feature extraction on training dataset with help of artificial neural network (ANN) employing a transferable regression-based QSAR model with r2>0.7 and acceptable internal and external validation criteria. The trained ANN is used to extract peptides from de novo helminth library. Derived peptides are filtered for their physiological and pharmaco-informatics perception such as instability index (<40), toxicity (IC50 ranging from 10-100μM potency), bioavailability (cell permeability), GRAVY (hydrophobicity index) etc. properties in silico. Peptides with suitable properties are studied for hot-spot binding to ACE2/Cathepsin L using molecular docking and the peptide fitting in the criteria of inhibitor to be filtered for stability with molecular dynamics simulation as per the guidelines. The target-drug models should be evaluated using a segmentation technique to scrutinize binding patches and scoring geometric complementary shapes or building a reproducible regression based QSAR model with published stable protein-peptide models and use this as a training set to predict the robustness of derived models. The peptide inhibitor with the prerequisite of ACE2/Cathepsin L inhibitor is synthesised and experimentally validated. Proposal will be evaluated on the strength of to build ANN/employ Machine learning on published stable protein-peptide models and use the training set to predict the robustness of derived models. The target-drug models to be evaluated using a segmentation technique to scrutinize binding patches and scoring geometric complementary shapes.