AlphaFold 3 developed by DeepMind and Isomorphic Labs, is an advanced artificial intelligence system that predicts the 3D structures of proteins and their interactions with other molecules, including DNA, RNA, ligands, and ions.
Availability and Restrictions
Versions
| Version | Pitzer | Ascend | Cardinal |
|---|---|---|---|
| 3.0.1 | X | X | X |
You can use module spider alphafold3 to view available modules for a given machine. Feel free to contact OSC Help if you need other versions for your work.
Access
AlphaFold 3 is available for all OSC users.
Publisher/Vendor/Repository and License Type
Copyright 2024 DeepMind Technologies Limited.
The AlphaFold 3 source code is licensed under the Creative Commons Attribution-Non-Commercial ShareAlike International License, Version 4.0 (CC-BY-NC-SA 4.0) (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at https://github.com/google-deepmind/alphafold3/blob/main/LICENSE.
Obtaining Model Parameters
The AlphaFold 3 model parameters are made available under the AlphaFold 3 Model Parameters Terms of Use (the "Terms"); you may not use these except in compliance with the Terms. You may obtain a copy of the Terms at https://github.com/google-deepmind/alphafold3/blob/main/WEIGHTS_TERMS_OF_USE.md.
Due to the Terms of Use, OSC will no longer maintain the model parameters in a central location. Users must download them by following the instructions at: https://github.com/google-deepmind/alphafold3?tab=readme-ov-file#obtaining-model-parameters.
Usage
Usage
Set-up
To load the default version of AlphaFold 3 module, use module load alphafold3.
Batch Usage
Below is the example batch script for an AlphaFold 3 job:
#!/bin/bash #SBATCH --nodes=1 #SBATCH --ntasks-per-node=8 #SBATCH --gpus-per-node=1 module reset module load alphafold3/3.0.1 run_alphafold.sh --model_dir=/path/to/your/model/parameters \ --output_dir=$(pwd -P)/output \ --json_path=2PV7.json
To get full-options list
run_alphafold.sh --helpshort
Note that the recommended hardware for AlphaFold 3 includes H100 (Cardinal) and A100 (Ascend). There are known issues with V100 (Pitzer), and additional parameters are required to run AlphaFold 3 on Pitzer, as referenced in the following links:
Using the example above, you need to modify the job script with the additional parameters, as shown below:
export APPTAINERENV_XLA_FLAGS="--xla_disable_hlo_passes=custom-kernel-fusion-rewriter"
run_alphafold.sh --model_dir=/path/to/your/model/parameters \
--output_dir=$(pwd -P)/output \
--json_path=2PV7.json \
--flash_attention_implementation=xla
Further Reading
- Online documentation is available on the AlphaFold 3 homepage.
- Notes on AlphaFold 3 output.
- Notes on citing AlphaFold 3.
