Show HN: Open-source GUI for editing 3D molecular structures with diffusion

DeepFold Protein released PATCHR-Studio, an open-source GUI for editing 3D molecular structures using diffusion-based inpainting to fill missing regions in proteins, DNA, RNA, and complexes. The tool is backend-agnostic, supports Boltz-2 and Protenix, and achieves a 99.4% connectivity pass rate, with a benchmark showing superior Cα RMSD compared to existing methods.

Structure inpainting and simulation-ready setup for proteins, DNA, RNA, and complexes Website https://patchr.deepfold.org/ | Atlas https://patchr.deepfold.org/atlas | Paper cite | PATCHR-Studio patchr-studio Download PATCHR-Studio: Windows https://github.com/DeepFoldProtein/patchr/releases/latest/download/patchr-studio-setup.exe · macOS Apple Silicon https://github.com/DeepFoldProtein/patchr/releases/latest/download/patchr-studio.dmg · Linux https://github.com/DeepFoldProtein/patchr/releases/latest/download/patchr-studio.AppImage 1KX3 : nucleosome histone-tail inpainting | 6GIS : PCNA + 50 bp DNA extension | 8GZR : NS3 polymerase + RNA reconstruction | 4ZLO : Kinase inpainting with ligand | Most experimental structures in the PDB have missing regions -- flexible loops, disordered terminals, unresolved sidechains. PATCHR fills them in using diffusion-based inpainting while keeping existing coordinates exactly as-is . Backend-agnostic -- supports Boltz-2 https://github.com/jwohlwend/boltz and Protenix https://github.com/bytedance/protenix - Works with proteins, DNA, RNA , and multi-chain complexes - 99.4% connectivity pass rate, from short loops to 600+ residue extensions PATCHR Atlas https://patchr.deepfold.org/atlas — large-scale inpainting of the PDB From every PDB complex with an internal missing region excluding very large structures , PATCHR inpainted the full set of 66,417 multimeric structures — all browsable and downloadable. Benchmark — 940 PDB40 structures with artificially introduced gaps mirroring real PDB missing-region statistics. Cα and all-atom RMSD computed over inpainted residues only. | Method / Configuration | Cα RMSD Å | All-atom RMSD Å | |---|---|---| PATCHR full, + LRD | 1.781 | 2.542 | | Boltz-2 + template conditioning | 4.647 | 5.510 | | Boltz-2 + template conditioning + steering threshold = 5.0 Å | 3.675 | 4.342 | | Boltz-2 + template conditioning + steering threshold = 2.0 Å | 3.397 | 4.081 | | Boltz-2 + template conditioning + steering threshold = 0.5 Å | 3.219 | 3.889 | | RFdiffusion2 all-atom | 9.188 | 10.199 | | RFdiffusion backbone-only | 2.043 | — | git clone https://github.com/DeepFoldProtein/patchr.git cd patchr && pip install -e . Mac conda create --name patchr python=3.12 llvmlite==0.44.0 numba==0.61.0 numpy==1.26.3 conda activate patchr git clone https://github.com/DeepFoldProtein/patchr.git cd patchr && pip install -e . export KMP DUPLICATE LIB OK=TRUE Docker ./scripts/docker-run.sh Run with all GPUs PATCHR GPU=0 ./scripts/docker-run.sh Select GPU Model weights are cached at ~/.boltz on the host override with BOLTZ CACHE . For Slurm clusters with Apptainer: sbatch scripts/slurm-run.sh 1. Generate a template from a PDB structure: patchr template 1TON all 2. Run inpainting: patchr predict examples/inpainting/1ton AB.yaml --out dir results The first run downloads the model checkpoint automatically to ~/.boltz/ . Template options patchr template 1CK4 all All polymer chains patchr template 4ZLO A,B --uniprot With UniProt sequence patchr template --input structure.cif A,B From local CIF patchr template 7EOQ all-copies Including duplicate copies patchr template 1BNA all -o my templates/ Custom output directory patchr template 7EOQ A --include-solvent Include solvent atoms patchr template 1CK4 all --assembly best Biological assembly patchr template 1CK4 all --relative-paths Use relative paths in YAML default: absolute Prediction options Single file patchr predict examples/inpainting/4zlo ABCD.yaml --out dir results --seed 42 patchr predict examples/inpainting/1ck4 AB.yaml --out dir results --diffusion samples 5 patchr predict examples/inpainting/1bna AB.yaml --out dir results --backend protenix patchr predict examples/inpainting/7eoq ABCDEFGHIJKLMN.yaml --out dir results --use msa server Bulk prediction — pass a directory of YAML files patchr predict my templates/ --out dir results patchr predict my templates/ --out dir results --backend protenix --seeds 42,101 Go directly from structure completion to MD simulation input: patchr predict input.yaml --out dir results --sim-ready gromacs patchr predict input.yaml --out dir results --sim-ready amber --ff amber14sb Standalone command patchr sim-ready prediction.cif --engine gromacs --ff charmm36m patchr sim-ready prediction.cif --engine openmm --padding 1.2 --ion-conc 0.15 PATCHR uses diffusion-based generation conditioned on your experimental structure as a rigid template: | Technique | What it does | | |---|---|---| | 1 | Template Conditioning | Anchors known coordinates at every diffusion step | | 2 | Synchronized Rigid Template Tracking | Keeps the template aligned with the evolving generation | | 3 | Local Refinement Denoising | Cleans up bond geometry at template-generation junctions | A desktop application providing a graphical interface to the full workflow, with no command line required. Download from the links above or the releases page https://github.com/DeepFoldProtein/patchr/releases . Beyond reconstructing missing regions, the interactive sequence editor supports residue-level edits applied directly on the structure and regenerated in a single inpainting run: Erase and regenerate — remove resolved residues and re-inpaint them Mutation — substitute a residue identity; the side chain is rebuilt by inpainting Post-translational modifications — introduce modified residues SEP, TPO, PTR, MLY, M3L Staged edits are listed and individually reversible prior to execution, and outputs are versioned for comparison across runs. Prediction, GPU queue status, and simulation-ready export are integrated. No GPU? Run the server on Google Colab https://colab.research.google.com/github/DeepFoldProtein/patchr/blob/main/colab server.ipynb for free and connect from PATCHR-Studio. patchr serve --model boltz2 --device-id 0 patchr serve --model protenix --port 8080 patchr serve --model all Beyond the headline RMSDs above, PATCHR also produces simulation-ready geometry: | Metric | Value | |---|---| | Backbone RMSD missing residues | 1.78 Å | | lDDT missing atoms | 98.6 | | Connectivity pass rate | 99.4% | Impact of Local Refinement Denoising LRD | Metric | With LRD | Without LRD | |---|---|---| | Structures with no issues | 99.4% | 87.4% | | Cα--Cα gaps 4.5--10 Å | 0.21% | 4.57% | | Peptide bond C--N issues | 0.85% | 15.43% | | Broken chains 10 Å | 0.32% | 0.74% | Accuracy by structural context | Secondary structure | RMSD Å | Solvent accessibility | RMSD Å | | |---|---|---|---|---| | Helix | 0.30 | Buried | 0.39 | | | Strand | 0.26 | Intermediate | 0.65 | | | Loop | 0.85 | Surface | 1.01 | Any model trained on the AlphaFold3 framework can be converted into an inpainting model through the PATCHR protocol. Currently implemented for Boltz-2 and Protenix only; extending to additional AF3-family backends is planned. PATCHR builds upon Boltz-2 https://github.com/jwohlwend/boltz by Passaro, Corso, Wohlwend et al. and Protenix https://github.com/bytedance/protenix by ByteDance. MIT -- free for academic and commercial use. @article{bae2025patchr, author = {Bae, Hanjin and Kim, Kunwoo and Yoo, Jejoong and Joo, Keehyoung}, title = {PATCHR-Studio: Template-conditioned diffusion-based molecular structure inpainting for Protein, RNA, and DNA complexes}, year = {2025} }