Protein Folding Python Packages

prospr

Prospr is a universal toolbox for protein structure prediction within the HP-model. The Python package is based on a C++ core, which gives Prospr its high performance. The C++ core is made available as a separate zip file to facilitate high-performance computing applications. The package comes with many prediction algorithms and datasets to use.

alphafold2-pytorch

To eventually become an unofficial Pytorch implementation / replication of Alphafold2, as details of the architecture get released

atlasfold

Trainable protein folding and co-folding models with a protein language model.

bionemo-ir

Easy, fast, and memory-efficient structure prediction inference

foldmetrics

Unified confidence metrics for structure-prediction models

invariant-point-attention

Implementation of Invariant Point Attention, used for coordinate refinement in the structure module of Alphafold2, as a standalone Pytorch module

molforge

A forge for protein workflows. One Python script, every tool in your stack.

equiformer-pytorch

Implementation of the Equiformer, SE3/E3 equivariant attention network that reaches new SOTA, and adopted for use by EquiFold for protein folding

refua-mcp

Model Context Protocol server exposing typed Refua tools for folding, affinity, design, and optional clinical and preclinical workflows.

clawcures

Python drug discovery toolkit for Boltz2 structure and affinity prediction, BoltzGen binder and antibody design, and TxGemma ADMET integration in one API.

refua

Python drug discovery toolkit for Boltz2 structure and affinity prediction, BoltzGen binder and antibody design, and TxGemma ADMET integration in one API.

foldpipe

Bounded-memory asynchronous streaming for native PyTorch/PyG molecular-data shards, with remote prefetch from Hugging Face and Google Drive.

intellifold

IntelliFold: A Controllable Foundation Model for General and Specialized Biomolecular Structure Prediction.

refua-clinical

Python drug discovery toolkit for Boltz2 structure and affinity prediction, BoltzGen binder and antibody design, and TxGemma ADMET integration in one API.

tr-rosetta-pytorch

Implementation of trRosetta and trDesign for Pytorch, made into a convenient package, for protein structure prediction and design

refua-data

Python drug discovery toolkit for Boltz2 structure and affinity prediction, BoltzGen binder and antibody design, and TxGemma ADMET integration in one API.

refua-deploy

Python drug discovery toolkit for Boltz2 structure and affinity prediction, BoltzGen binder and antibody design, and TxGemma ADMET integration in one API.

refua-regulatory

Python drug discovery toolkit for Boltz2 structure and affinity prediction, BoltzGen binder and antibody design, and TxGemma ADMET integration in one API.

frustration-gpu

Pure-PyTorch LAMMPS-AWSEM frustration analysis. 14-53x faster than frustrapy on a single GPU. Byte-comparable to frustratometeR.

refua-studio

Python drug discovery toolkit for Boltz2 structure and affinity prediction, BoltzGen binder and antibody design, and TxGemma ADMET integration in one API.

refua-bench

Python drug discovery toolkit for Boltz2 structure and affinity prediction, BoltzGen binder and antibody design, and TxGemma ADMET integration in one API.

refua-wetlab

Python drug discovery toolkit for Boltz2 structure and affinity prediction, BoltzGen binder and antibody design, and TxGemma ADMET integration in one API.

triangle-multiplicative-module

Implementation of the Triangle Multiplicative module, used in Alphafold2 as an efficient way to mix rows or columns of a 2d feature map, as a standalone package for Pytorch

jaxfvld

Small, fast, fully differentiable antibody Fv structure predictor in JAX. ~0.76M parameters, matches IgFold on CDR-H3, folds an Fv in ~0.8s on CPU.