A collection of Apptainer (.def) and Docker (Dockerfile) container definitions for GPU-accelerated protein design and structure prediction tools. Each subfolder is self-contained with a build definition, run instructions, and license.
| Subfolder | Tool | Task |
|---|---|---|
| AlphaFold3/ | AlphaFold 3 | Structure prediction (proteins, ligands, nucleic acids) |
| bindcraft/ | BindCraft | De novo protein binder design (AF2-based) |
| boltz2/ | Boltz-2 | Structure prediction + binding affinity |
| boltzgen/ | BoltzGen | Generative design of proteins, peptides, and binders |
| CARBonAra/ | CARBonAra | Inverse folding / sequence design |
| Chai/ | Chai-1 | Structure prediction (proteins, ligands, DNA, RNA) |
| esm/ | ESM2 | Protein language model embeddings and folding |
| LigandMPNN/ | LigandMPNN | Inverse folding with ligand/DNA context |
| RFDiffusion3/ | RFdiffusion3 | Diffusion-based all-atom biomolecular design |
See LINK_COLLECTION.md for online demos and further tools not yet containerised here.
If you don't want to build the containers yourself, feel free to reach out and I can send you the pre-built .sif files individually.
Build:
apptainer build --force ./<tool>.sif ./<tool>.def
Run (GPU passthrough, bind-mount a local data directory):
apptainer exec --nv --bind "$(pwd)":/data <tool>.sif <command>
See the README.md in each subfolder for the tool-specific command.
Build:
docker build -t <tool>:latest .
Convert to Singularity/Apptainer if needed:
singularity build <tool>.sif docker-daemon://<tool>:latest
- All containers require an NVIDIA GPU (
--nv/--gpus all). - AlphaFold 3 weights must be requested separately — see AlphaFold3/README.md.
- BoltzGen and ESM2 download model weights automatically on first run.
This repository contains Apptainer/Docker container definitions that package third-party software. Each tool is subject to its own license; see the LICENSE.md file in the respective subfolder for the full license text.
| Tool | License | Source |
|---|---|---|
| AlphaFold 3 | CC BY-NC-SA 4.0 | https://github.com/google-deepmind/alphafold3 |
| BindCraft | MIT | https://github.com/martinpacesa/BindCraft |
| Boltz-2 | MIT | https://github.com/jwohlwend/boltz |
| BoltzGen | MIT | https://github.com/hstark/boltzgen |
| Chai-1 | Apache 2.0 | https://github.com/chaidiscovery/chai-lab |
| CARBonAra | CC BY-NC-SA 4.0 | https://github.com/ideaslabut/CARBonAra |
| ESM2 | MIT | https://github.com/facebookresearch/esm |
| LigandMPNN | MIT | https://github.com/dauparas/LigandMPNN |
| RFdiffusion3 | BSD 3-Clause | https://github.com/RosettaCommons/RFdiffusion |
The container build scripts in this repository are provided as-is. Users are responsible for complying with the licenses of all bundled software, including any non-commercial restrictions (e.g. AlphaFold 3 and CARBonAra are CC BY-NC-SA 4.0).