AlphaFold2

Last Updated: March 3, 2025


Module

Module Name Version
alphafold 2.3.2

This module is not required for execution. The module help command can be used to display information about a sample job script.


Overview

AlphaFold2 is a software developed by DeepMind that uses machine learning techniques to predict the three-dimensional structure of proteins and other molecules.


How to use

Since it takes time to access the database when AlphaFold2 is executed, the necessary database files are prepared on the high-speed storage. A sample job script to run AlphaFold2 using the prepared database file is provided below. Copy the file to your home directory and modify it as necessary. Since it seems that using multiple GPUs does not seem to improve performance, the job is designed to use only one GPU.

  • For monomer model
    • /home/center/app/alphafold/v2.3.2/samples/run_monomer.sh
  • For multimer model
    • /home/center/app/alphafold/v2.3.2/samples/run_multimer.sh

The scripts that are called at the end of these scripts and actually executed are as follows

  • /home/center/app/alphafold/v2.3.2/alphafold/bin/alphafold

If the execution options do not match your expectation, please copy and modify this script. The path to the database on the fast storage (/fast/center/AFDB/v2.3) is also described in this file. The contents of this file may be updated without notice. If you want to replace some of the databases, modify the arguments at the end of the script and execute it.