Generative Chemistry & Drug Discovery🧪💊
Collection
This collection explores generative and predictive modeling for molecular data, combining sequence-based generation (VAE/RNN) with graph model. • 2 items • Updated
Multi-Head Graph Isomorphism Network (GINE) for Molecular Property Prediction
This is a Multi-Head Graph Isomorphism Network (GINE) model designed for predicting molecular properties such as lipophilicity, molecular weight, hydrogen bond donor count, and hydrogen bond acceptor count from SMILES strings. The model takes a SMILES string as input, converts it into a graph representation, and outputs the predicted properties. Training data from ChemBL library
Full project file at https://github.com/teohyc/drug_agent
Usage
from prop_gnn_infer import predict_mol
from prop_gnn_model import MoleculeGINE
#change to your test SMILES strings
print(predict_mol(test_smiles=["O=C1N=C2SCCN2C(=O)C1Cc1ccc(Cl)cc1", "C[C@@H]1C[C@H]2[C@@H]3CCC4=CC(=O)C=C[C@]4(C)[C@@]3(F)[C@@H](O)C[C@]2(C)[C@@]1(C)C(=O)CO"]))
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support