DrugHIVE: Structure-based drug design with a deep hierarchical generative model
[PNAS 2025] Code of "Manifold-Constrained Nucleus-Level Denoising Diffusion Model for Structure-Based Drug Design".
Official implementation of "BInD: Bond and Interaction-Generating Diffusion Model for Multi-Objective Structure-Based Drug Design" (Advanced Science)
Research repository for diffusion based structure based drug design
Open-source SBDD toolkit for standardized import of binding poses, restrained OpenMM minimization, pose QC, protein-ligand interaction analysis, ligand strain analysis, MD simulation, and integrated design prioritization. Supports single receptors, but also multichain receptors, molecular glues, and PROTAC ternary complexes.
TAGMol: Target-Aware Gradient-guided Molecule Generation (ICML'24 ML4LMS Workshop)
UV-first benchmark for protein-ligand docking with reproducible annotation, pocket similarity, and HPC workflows.
Official implementation of the pre-print "Coupled Fragment-Based Generative Modeling with Stochastic Interpolants" by Tuan Le, Yanfei Guan, Djork-Arné Clevert and Kristof T. Schütt.
ProteaFlow — de novo protein binder and small molecule design using 10 generative tools. Automated validation, scoring, and ranking pipeline. Intuitive web interface for job submission, live monitoring, and interactive result exploration — no coding required.
Structure-based paralog selectivity counter-screen for the immuno-oncology target ENPP1 against ENPP2 (autotaxin) and ENPP3 — ranked by cross-paralog margin, not raw affinity.
LigandForge is a modular ligand preparation and virtual screening pipeline for drug discovery that integrates ADMET filtering, 3D conformer generation, Meeko-based ligand preparation, and AutoDock Vina docking—with full provenance tracking, checkpointing, and clean restarts.
Computational tool for automated analysis of protein–ligand interactions, docking results, and ligand structural descriptors in structure-based drug discovery.
Validated computational drug-discovery pipeline for KRAS in pancreatic ductal adenocarcinoma (PDAC). Four validated targets (G12C/G12D/G12V/G12R), 518,662 gated compounds, three virtual screens, MM-GBSA rescoring, and a self-audit of a generative campaign whose molecules could not be synthesised. All negative results released.
This package facilitates molecular docking simulations aimed at analyzing interactions between a target biological system and a collection of potential drug molecules. By leveraging computational algorithms, it ranks these molecules based on docking scores and interaction energies, providing insights into their suitability as drug candidates.
Benchmarking and repairing metal coordination failures in pocket-conditioned 3D generative models for structure-based drug design (SBDD).
Agentic drug discovery workflows on fully open-source tooling: RDKit, AutoDock Vina/smina, fpocket, ChEMBL. MCP tool servers + Claude agents, with a retrospective validation harness.
Atomwise — independent third-party profile of a public API surface, by API Evangelist. Atomwise pioneered the use of deep convolutional neural networks for structure-based drug discovery, with its AtomNet platform applying AI to virtual screening of small molecules against protein targets.
ChemApp is a Streamlit-based platform for virtual screening and early-stage drug discovery, integrating machine learning, molecular generation, fingerprinting, drug-likeness assessment, molecular docking, and interactive visualization in a unified workflow.
Probe-based spherical interaction fields for pocket-conditioned molecule generation.
Portfolio of mini-projects for upskilling in ML applied to Cheminformatics and Computational Chemistry
To associate your repository with the structure-based-drug-design topic, visit your repo's landing page and select "manage topics."