online-chem

online-chem generates novel small molecules as potential Mdmx (Mdm2-associated X protein) inhibitors for oncology-focused drug discovery.

Key Features:

• Focused Library Generator: Generates de novo molecular entities and is trained on the ChEMBL database of bioactive, drug-like molecules.
• Transfer Learning for Targeted Generation: Uses transfer learning to adapt the generative model toward compounds targeting Mdmx (Mdm2-associated X protein).
• Integration with Computational Chemistry Tools: Applies Lilly medicinal chemistry filters and uses OpenBabel to convert structures into PDBQT format for downstream analysis.
• Molecular Docking and QSAR Modeling: Employs molecular docking alongside a QSAR IC50 model to predict binding affinity and inhibitory potency.
• Pharmacophore Screening and Molecular Dynamics Simulations: Performs pharmacophore screening and molecular dynamics (MD) simulations to evaluate pharmacophoric features and ligand stability.
• Identification of Promising Hits: Identifies five Mdmx inhibitor candidates with predicted binding free energies and IC50 values comparable or superior to known inhibitors.

Scientific Applications:

• Medicinal Chemistry: Supports in silico generation and prioritization of candidate small molecules for medicinal chemistry lead discovery.
• Oncology Research: Aids discovery of anticancer compounds targeting Mdmx, a protein implicated in cancer progression.

Methodology:

Computational methods explicitly include training on ChEMBL, transfer learning, de novo molecule generation, application of Lilly medicinal chemistry filters, OpenBabel conversion to PDBQT, molecular docking, QSAR IC50 modeling, pharmacophore screening, and molecular dynamics (MD) simulations.