AspirinCode/awesome-AI4MolConformation-MD
List of molecules (small molecules, RNA, peptide, protein, enzymes, antibody, and PPIs) conformations and molecular dynamics (force fields) using generative artificial intelligence and deep learning
What it solves
Researchers working on molecular structures and simulations often struggle to find relevant papers, datasets, and tools that apply modern AI techniques. This project is a curated list that organizes that scattered information into one place. It collects resources on predicting and simulating the 3D shapes (conformations) of molecules such as proteins, RNA, small molecules, and peptides, as well as on molecular dynamics (MD) simulations that use force fields. The list covers both traditional physics-based methods and newer deep-learning and generative AI approaches.
How it works
The project is a living repository of links and citations, grouped under a clear menu. It starts with general reviews, then lists datasets and software packages. The main sections cover molecular dynamics, classical and neural force fields, MD engines, trajectory analysis, visualization tools, and specialized topics like coarse-grained MD, neural network potentials, and reactive force fields. It also organizes resources by the type of molecule (e.g., small molecule, peptide, protein, enzyme, antibody) and by the method used (e.g., AlphaFold-based, VAE-based, diffusion-based, flow-based, GAN-based, transformer-based). Each entry includes the title, authors, year, and a link to the publication or code/data. The list is explicitly marked as "Updating ...", meaning it is intended to grow over time.
Who it's for
This is aimed at computational chemists, biophysicists, and machine learning researchers who study molecular structure and dynamics. It is also useful for drug discovery scientists interested in protein-ligand interactions, antibody conformations, or free energy calculations. The project helps people quickly find benchmarks, training datasets, simulation packages, and force field parameters for their own work.
Highlights
- Covers a broad range of molecule types: small molecules, RNA, peptides, proteins, enzymes, antibodies, and protein-protein interactions.
- Includes both classical molecular dynamics and AI-enhanced methods such as neural network potentials and generative models.
- Organizes resources by technique (diffusion, GAN, VAE, transformer, etc.) and by application (conformational dynamics, free energy perturbation, etc.).
- Curated lists of datasets and packages such as DynaRepo, mdCATH, nablaDFT, and MMolearn.
- Provides direct links to papers, code repositories, and data downloads.
Related
- Project
- Project
- Project
- Project