QPoweredCompound2DeNovoDrugPropMax

QPoweredCompound2DeNovoDrugPropMax: Deep Learning and Quantum-Enhanced Network Pharmacology Platform

QPoweredCompound2DeNovoDrugPropMax predicts compound–target interactions and optimizes drug-like properties using deep learning and quantum-classical hybrid neural networks trained on BindingDB interaction data.

Key Features:

• Deep Learning-Based Network Pharmacology: Trains neural networks on BindingDB compound–drug target interaction data to classify PubChem compounds by predicted target interactions represented by RCSB PDB IDs.
• PubChem CID-Based Prediction: Accepts PubChem Compound IDs (CIDs) and performs multi-class classification to predict associated biological targets and activities.
• Drug-Likeness Optimization: Applies a deep learning-based structure optimization protocol to enhance physicochemical and drug-like properties of candidate compounds.
• Protein–Ligand Interaction Profiling: Automates in silico modeling to generate protein–ligand interaction profiles and analyze predicted binding mechanisms and affinities.
• Quantum-Classical Hybrid Modeling: Integrates quantum layers into classical neural networks using the PennyLane interface to quantum hardware to improve predictive performance.

Scientific Applications:

• Target Interaction Prediction: Identifies likely interactions between novel compounds and biological drug targets using BindingDB and RCSB PDB-referenced data.
• Drug Development Optimization: Refines candidate compounds for improved drug-likeness during early-stage therapeutic development.
• Protein–Ligand Interaction Analysis: Characterizes predicted binding interactions to support mechanistic studies and rational drug design.

Methodology:

Trains deep learning neural networks on BindingDB compound–target interaction datasets to perform multi-class classification of PubChem CIDs by predicted RCSB PDB target associations. Incorporates quantum-classical hybrid architectures via PennyLane to augment classical neural networks for compound–target prediction and structure optimization.