Tutorial on running Colbafold for AlphaFold2 predictions using a Linux Cluster with HTCondor. A description blog (AlphaFold2 with ColabFold in Container) was written describing the process, but scripts can be found in the complete tutorial as well as plain text files below.
Text scripts and files can be copied from the tutorial, but for convenience are provided as individual files below. Each file has an extra .txt at the end to be acceptable by the web server. and should be removed. The prefix `bcc_` and `chtc_` were added to distinguish the files to be run on different servers as detailed in the tutorial.
All files in single zip file: runfiles.zip
Individual files:
– bcc_runaf.sh.txt
– bcc_runaf.sub.txt
– chtc_dl.sh.txt
– chtc_download_weights.sub.txt
– chtc_runaf.sh.txt
– chtc_runaf.sub.txt
– hemoglobin-colab.fa.txt
– test.fa.txt
Large-scale batch predictions
The set-up described in this tutorial is suitable for a few runs, but for very large batch runs other resources should be used.In particular, this tutorial made the specific effort to provide the necessary information and scripts to run on a cluster of Linux computer with HTCondor. The resources below assume a direct use of the container.
Using Colabfold this blog titled Dockerized Colabfold for large-scale batch predictions [Archived – 2024-11-20] provides detailed information to add a large database.
This blog titled How to Install ColabFold & Run AI Protein Folding Locally [Archived – 2024-11-20] provides a detailed explanation to install Colabfold on one’s laptop/desktop with 2 alternate github repositories.
Both provide details on how Colabfold helps run a prediction faster than a “native” installation.