Quick adaptation to run RaSP using CPU and NextFlow

This repo can be used to quickly install and run RaSP on multiple protein structures using CPU.

Adapted from RaSP Colab Notebook RaSP Colab Notebook and RaSP GitHub repo. All credits to Rapid protein stability prediction using deep learning representations.

Install requirements

1. Install conda
2. Install nextflow (version 23.04.3 was used to test the run)
3. Clone this repository

Run RaSP

By running the following command, it will predict the stability change upon mutations for every protein included in the --indir directory.

nextflow run  main.nf --indir structures/test_structures/ --outdir output/predictions/

The pipeline was used to run RaSP on the entire human and mouse proteomes derived from AlphaFold 2 (version 4) data available on AlphaFold Database. With the specified parameters --cores 1, --memory 5.GB, and --max_running 100, the pipeline completed the predictions within approximately one week for each proteome.