ChemProp: machine learning based property prediction of molecules

Blog#5

ChemProp can be used to estimate the properties of molecules by utilising a Message Passing Neural Network (MPNN). To make predictions, an MPNN must first be trained on a dataset of molecules with known property values. Once trained, the MPNN can be used to predict similar properties in any new molecule.

Here is the link:http://chemprop.csail.mit.edu

Till now it has provided models for predicting that molecules can inhibit the growth of bacteria ( antibiotic property), 3Cl pro/mpro of sars-cov.

The SARS balanced model is for more exhaustive prediction.

AMU SARS-CoV-2 in vitro is for the  estimation of probability that molecule with inhibit the virus replication in vitro)