kexinhuang12345/DeepPurpose
A Deep Learning Toolkit for DTI, Drug Property, PPI, DDI, Protein Function Prediction (Bioinformatics)
Combines 15+ molecular encoding strategies—from RDKit fingerprints and CNNs to graph neural networks (GCN, GIN, AttentiveFP via DGL)—with 50+ model variants for flexible compound-protein representation learning. Built on PyTorch with integrated support for cold-start evaluation, hyperparameter tuning via Ax/Bayesian optimization, and automatic task detection (binary classification vs. regression). Seamlessly integrates with the TDC therapeutics dataset loader and provides pre-trained checkpoints on BindingDB, DAVIS, and KIBA datasets for rapid prototyping.
1,134 stars. No commits in the last 6 months. Available on PyPI.
Stars
1,134
Forks
299
Language
Jupyter Notebook
License
BSD-3-Clause
Category
Last pushed
Jun 10, 2024
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/kexinhuang12345/DeepPurpose"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Related frameworks
BioinfoMachineLearning/PoseBench
Comprehensive benchmarking of protein-ligand structure prediction methods. (Nature Machine Intelligence)
pritampanda15/PandaDock
PandaDock: Physics based Molecular Docking with GNN Scoring
kamerlinlab/KIF
KIF - Key Interactions Finder. A python package to identify the key molecular interactions that...
maranasgroup/CatPred
Machine Learning models for in vitro enzyme kinetic parameter prediction
BioinfoMachineLearning/FlowDock
A geometric flow matching model for generative protein-ligand docking and affinity prediction....