RFdiffusion

This is the Kuhlman Lab fork of RFdiffusion.

RFdiffusion is an open source method for structure generation, with or without conditional information (a motif, target etc). It can perform a whole range of protein design challenges as outlined in the RFdiffusion paper.

All of the functionality of the original repo is preserved (refer to the original README for detailed explanations). A few additional changes have been made in this repo:

• Symmetry: Random chain lengths can now be used no matter what symmetry order.
• New potentials: 1) loop_contacts for biasing the distance between two residues, 2) hetero_olig for biasing number of inter- and intra-chain contacts, 3) binder_RMSD for biasing binder topologies towards a specific shape based on RMSD, 4) res_pair_constraints for biasing binder topologies towards a specific shape based on residue pair distances.
• Utilities for interacting with EvoPro.

Installation

We provide an alternative installation process for RFdiffusion.

1. Clone the repo:

git clone https://github.com/Kuhlman-Lab/RFdiffusion.git

2. Download the model parameters:

cd RFdiffusion
mkdir models; cd models
wget http://files.ipd.uw.edu/pub/RFdiffusion/6f5902ac237024bdd0c176cb93063dc4/Base_ckpt.pt
wget http://files.ipd.uw.edu/pub/RFdiffusion/e29311f6f1bf1af907f9ef9f44b8328b/Complex_base_ckpt.pt
wget http://files.ipd.uw.edu/pub/RFdiffusion/60f09a193fb5e5ccdc4980417708dbab/Complex_Fold_base_ckpt.pt
wget http://files.ipd.uw.edu/pub/RFdiffusion/74f51cfb8b440f50d70878e05361d8f0/InpaintSeq_ckpt.pt
wget http://files.ipd.uw.edu/pub/RFdiffusion/76d00716416567174cdb7ca96e208296/InpaintSeq_Fold_ckpt.pt
wget http://files.ipd.uw.edu/pub/RFdiffusion/5532d2e1f3a4738decd58b19d633b3c3/ActiveSite_ckpt.pt
wget http://files.ipd.uw.edu/pub/RFdiffusion/12fc204edeae5b57713c5ad7dcb97d39/Base_epoch8_ckpt.pt
wget http://files.ipd.uw.edu/pub/RFdiffusion/f572d396fae9206628714fb2ce00f72e/Complex_beta_ckpt.pt

3. Create a new environment based on the yaml file (rfdiff_env.yaml):

cd ../env
conda env create -n rfdiff -f rfdiff_env.yaml

4. Activate the new environment:

conda activate rfdiff

5. Install dllogger:

pip install "dllogger @ git+https://github.com/NVIDIA/dllogger.git"

6. Install se3-transformer:

cd SE3Transformer
python setup.py install

7. Install rfdiffusion:

cd ../..
pip install -e .

Note: It should also be possible to follow the instructions in the original repo, but only difference is to clone this repo (git clone https://github.com/Kuhlman-Lab/RFdiffusion.git) rather than the original.