LanceKnight / MolKGNN
MolKGNN is a deep learning model for predicting biological activity or molecular properties. It features in 1. SE(3)-invariance 2. conformation-invariance 3. interpretability. MolKGNN uses a novel molecular convolution to leverage the similarity of molecular neighborhood and kernels. It shows superior results in realistic drug discovery dataset…
☆12Updated 10 months ago
Related projects ⓘ
Alternatives and complementary repositories for MolKGNN
- ACGCN: Graph Convolutional Networks for Activity Cliff Prediction Between Matched Molecular Pairs (Park et al., 2022)☆19Updated last year
- Implementation of iMolCLR: "Improving Molecular Contrastive Learning via Faulty Negative Mitigation and Decomposed Fragment Contrast" in …☆17Updated 2 years ago
- Chemprop benchmarking scripts and data for v1☆25Updated 5 months ago
- ☆14Updated last year
- ☆22Updated 5 months ago
- EdgeSHAPer: Bond-Centric Shapley Value-Based Explanation Method for Graph Neural Networks☆24Updated last year
- ProCare: A Point Cloud Registration Approach to Align Protein Cavities☆29Updated last year
- Molecule Graph Generation using Graph Convolutional Networks-based Variational Graph AutoEncoders (VGAE) in PyTorch☆21Updated 2 weeks ago
- Python API for Pharmer☆11Updated 5 years ago
- Bias-controlled 3D generative framework for structure-based ligand design☆17Updated 2 years ago
- ☆14Updated 2 years ago
- ☆11Updated last month
- Supporting models and data to doi 10.1021/acs.jcim.1c01163☆15Updated 2 years ago
- ☆11Updated 2 years ago
- Official Github for "Molecular generative model via retrosynthetically prepared chemical building block assembly" (Advanced Science)☆28Updated last month
- ☆23Updated 5 months ago
- ☆22Updated last week
- COMA: Efficient Structure-constrained Molecular Generation using Contractive and Margin losses☆18Updated last year
- Code and datasets from the publication https://doi.org/10.1186/s13321-023-00787-9☆19Updated 7 months ago
- Source code accompanying the 'MF-PCBA: Multi-fidelity high-throughput screening benchmarks for drug discovery and machine learning' paper☆24Updated last year
- A method for ranking fragments by how much novel information they give about protein targets in fragment screens. When using the results …☆9Updated 2 years ago
- RJT-RL: De novo molecular design using a Reversible Junction Tree and Reinforcement Learning☆23Updated 2 years ago
- ☆20Updated 3 years ago
- ☆12Updated 3 years ago
- Code for paper "Human-in-the-Loop Assisted de Novo Molecular Design".☆22Updated last year
- MolDesigner: Interactive Design of Efficacious Drugs with Deep Learning (NeurIPS 2020 Demo)☆14Updated 3 years ago
- Code for "De novo molecular design with chemical language models"☆12Updated 3 years ago
- ☆10Updated last year
- Official code for the publication "HyFactor: Hydrogen-count labelled graph-based defactorization Autoencoder".☆17Updated last year
- Semi-supervised Junction Tree Variational Autoencoder for jointly trained property prediction and molecule structure generation. (AAAI 23…☆11Updated last year