Merge branch 'RosettaCommons:main' into main

Author: geraseva

.github/workflows/main.yml

@@ -0,0 +1,95 @@
+# This workflow will install Python dependencies, run tests and lint with a single version of Python
+# For more information see: https://docs.github.com/en/actions/automating-builds-and-tests/building-and-testing-python
+
+name: Python application
+
+on:
+  push:
+    branches: [ "main", "github-ci" ]
+  pull_request:
+    branches: [ "main", "github-ci" ]
+
+permissions:
+  contents: read
+
+jobs:
+  ppi-scaffolds-test:
+
+    #runs-on: ubuntu-latest
+    runs-on: ubuntu-24.04
+    #container: nvcr.io/nvidia/cuda:12.9.1-cudnn-runtime-ubuntu20.04
+    container: nvcr.io/nvidia/cuda:11.6.2-cudnn8-runtime-ubuntu20.04
+
+    steps:
+    - uses: actions/checkout@v4
+    - name: Set up Python 3.9
+      uses: actions/setup-python@v3
+      with:
+        python-version: "3.9"
+    # - name: Install dependencies
+    #   run: |
+    #     python -m pip install --upgrade pip
+    #     #pip install flake8 pytest
+    #     pip install pytest
+    #     if [ -f requirements.txt ]; then pip install -r requirements.txt; fi
+    # - name: Lint with flake8
+    #   run: |
+    #     # stop the build if there are Python syntax errors or undefined names
+    #     flake8 . --count --select=E9,F63,F7,F82 --show-source --statistics
+    #     # exit-zero treats all errors as warnings. The GitHub editor is 127 chars wide
+    #     flake8 . --count --exit-zero --max-complexity=10 --max-line-length=127 --statistics
+
+    - name: Install dependencies
+      run: |
+        apt-get update -qq && apt-get install -y --no-install-recommends \
+        build-essential \
+        git \
+        curl \
+        wget \
+        ca-certificates
+
+        curl -LsSf https://astral.sh/uv/install.sh | sh
+        . $HOME/.local/bin/env bash
+
+        uv python install 3.9
+        uv venv
+        uv pip install --no-cache-dir -q \
+          dgl==1.0.2+cu116 -f https://data.dgl.ai/wheels/cu116/repo.html \
+          torch==1.12.1+cu116 --extra-index-url https://download.pytorch.org/whl/cu116 \
+          e3nn==0.3.3 \
+          wandb==0.12.0 \
+          pynvml==11.0.0 \
+          git+https://github.com/NVIDIA/dllogger#egg=dllogger \
+          decorator==5.1.0 \
+          hydra-core==1.3.2 \
+          pyrsistent==0.19.3 \
+          pytest
+
+        uv pip install --no-cache-dir env/SE3Transformer
+        uv pip install --no-cache-dir -e . --no-deps
+        rm -rf ~/.cache # /app/RFdiffusion/tests
+
+    - name: Download weights
+      run: |
+        mkdir models
+        wget -q -O models/Base_ckpt.pt http://files.ipd.uw.edu/pub/RFdiffusion/6f5902ac237024bdd0c176cb93063dc4/Base_ckpt.pt
+        wget -q -O models/Complex_base_ckpt.pt http://files.ipd.uw.edu/pub/RFdiffusion/e29311f6f1bf1af907f9ef9f44b8328b/Complex_base_ckpt.pt
+        wget -q -O models/Complex_Fold_base_ckpt.pt http://files.ipd.uw.edu/pub/RFdiffusion/60f09a193fb5e5ccdc4980417708dbab/Complex_Fold_base_ckpt.pt
+        wget -q -O models/InpaintSeq_ckpt.pt http://files.ipd.uw.edu/pub/RFdiffusion/74f51cfb8b440f50d70878e05361d8f0/InpaintSeq_ckpt.pt
+        wget -q -O models/InpaintSeq_Fold_ckpt.pt http://files.ipd.uw.edu/pub/RFdiffusion/76d00716416567174cdb7ca96e208296/InpaintSeq_Fold_ckpt.pt
+        wget -q -O models/ActiveSite_ckpt.pt http://files.ipd.uw.edu/pub/RFdiffusion/5532d2e1f3a4738decd58b19d633b3c3/ActiveSite_ckpt.pt
+        wget -q -O models/Base_epoch8_ckpt.pt http://files.ipd.uw.edu/pub/RFdiffusion/12fc204edeae5b57713c5ad7dcb97d39/Base_epoch8_ckpt.pt
+        #optional
+        wget -q -O models/Complex_beta_ckpt.pt http://files.ipd.uw.edu/pub/RFdiffusion/f572d396fae9206628714fb2ce00f72e/Complex_beta_ckpt.pt
+        #original structure prediction weights
+        wget -q -O models/RF_structure_prediction_weights.pt http://files.ipd.uw.edu/pub/RFdiffusion/1befcb9b28e2f778f53d47f18b7597fa/RF_structure_prediction_weights.pt
+
+    - name: Setup and Run ppi_scaffolds tests
+      run: |
+        tar -xvf examples/ppi_scaffolds_subset.tar.gz -C examples
+        cd tests && uv run python test_diffusion.py
+
+
+    # - name: Test with pytest
+    #   run: |
+    #     pytest

README.md

@@ -45,6 +45,7 @@ RFdiffusion is an open source method for structure generation, with or without c
     - [Generation of Symmetric Oligomers](#generation-of-symmetric-oligomers)
     - [Using Auxiliary Potentials](#using-auxiliary-potentials)
     - [Symmetric Motif Scaffolding.](#symmetric-motif-scaffolding)
+    - [RFpeptides macrocycle design](#macrocyclic-peptide-design-with-rfpeptides)
     - [A Note on Model Weights](#a-note-on-model-weights)
     - [Things you might want to play with at inference time](#things-you-might-want-to-play-with-at-inference-time)
     - [Understanding the output files](#understanding-the-output-files)
@@ -466,6 +467,49 @@ Note that the contigs should specify something that is precisely symmetric. Thin
 
 ---
 
+### Macrocyclic peptide design with RFpeptides
+<img src="./img/rfpeptides_fig1.png" alt="alt text" width="400px" align="right"/>
+We have recently published the RFpeptides protocol for using RFdiffusion to design macrocyclic peptides that bind target proteins with atomic accuracy (Rettie, Juergens, Adebomi, et al., 2025). In this section we briefly outline how to run this inference protocol. We have added two examples for running macrocycle design with the RFpeptides protocol. One for monomeric design, and one for binder design.
+
+```
+examples/design_macrocyclic_monomer.sh
+examples/design_macrocyclic_binder.sh
+```
+#### RFpeptides binder design
+<img src="./img/rfpeptides_binder.png" alt="alt text" width="1100" align="center"/>
+
+To design a macrocyclic peptide to bind a target, the flags needed are very similar to classic binder design, but with two additional flags: 
+```
+#!/bin/bash 
+
+prefix=./outputs/diffused_binder_cyclic2
+
+# Note that the indices in this pdb file have been 
+# shifted by +2 in chain A relative to pdbID 7zkr.
+pdb='./input_pdbs/7zkr_GABARAP.pdb'
+
+num_designs=10
+script="../scripts/run_inference.py"
+$script --config-name base \
+inference.output_prefix=$prefix \
+inference.num_designs=$num_designs \
+'contigmap.contigs=[12-18 A3-117/0]' \
+inference.input_pdb=$pdb \
+inference.cyclic=True \
+diffuser.T=50 \
+inference.cyc_chains='a' \
+ppi.hotspot_res=[\'A51\',\'A52\',\'A50\',\'A48\',\'A62\',\'A65\'] \
+```
+
+The new flags are `inference.cyclic=True` and `inference.cyc_chains`. Yes, they are somewhat redundant. 
+
+`inference.cyclic` simply notifies the program that the user would like to design at least one macrocycle, and `inference.cyc_chains` is just a string containing the letter of every chain you would like to design as a cyclic peptide. In the example above, only chain `A` (`inference.cyc_chains='a'`) is cyclized, but one could do `inference.cyc_chains='abcd'` if they so desired (and the contigs was compatible with this, which the above one is not). 
+
+#### RFpeptides monomer design
+For monomer design, you can simply adjust the contigs to only contain a single generated chain e.g., `contigmap.contigs=[12-18]`, keep the `inference.cyclic=True` and `inference.cyc_chains='a'`, and you're off to the races making monomers.
+
+---
+
 ### A Note on Model Weights
 
 Because of everything we want diffusion to be able to do, there is not *One Model To Rule Them All*. E.g., if you want to run with secondary structure conditioning, this requires a different model than if you don't. Under the hood, we take care of most of this by default - we parse your input and work out the most appropriate checkpoint.

config/inference/base.yaml

@@ -21,6 +21,8 @@ inference:
   trb_save_ckpt_path: null
   schedule_directory_path: null
   model_directory_path: null
+  cyclic: False 
+  cyc_chains: 'a'
 
 contigmap:
   contigs: null

examples/design_macrocyclic_binder.sh

@@ -0,0 +1,19 @@
+#!/bin/bash
+
+prefix=./outputs/diffused_binder_cyclic2
+
+# Note that the indices in this pdb file have been 
+# shifted by +2 in chain A relative to pdbID 7zkr.
+pdb='./input_pdbs/7zkr_GABARAP.pdb'
+
+num_designs=10
+script="../scripts/run_inference.py"
+$script --config-name base \
+inference.output_prefix=$prefix \
+inference.num_designs=$num_designs \
+'contigmap.contigs=[12-18 A3-117/0]' \
+inference.input_pdb=$pdb \
+inference.cyclic=True \
+diffuser.T=50 \
+inference.cyc_chains='a' \
+ppi.hotspot_res=[\'A51\',\'A52\',\'A50\',\'A48\',\'A62\',\'A65\'] \

examples/design_macrocyclic_monomer.sh

@@ -0,0 +1,17 @@
+#!/bin/bash 
+
+prefix=./outputs/uncond_cycpep
+# Note that the indices in this pdb file have been 
+# shifted by +2 in chain A relative to pdbID 7zkr.
+pdb='./input_pdbs/7zkr_GABARAP.pdb'
+
+num_designs=10
+script="../scripts/run_inference.py"
+$script --config-name base \
+inference.output_prefix=$prefix \
+inference.num_designs=$num_designs \
+'contigmap.contigs=[12-18]' \
+inference.input_pdb=$pdb \
+inference.cyclic=True \
+diffuser.T=50 \
+inference.cyc_chains='a'