MMseqs2

MMseqs2 (Many-against-Many sequence searching) is a software suite for very fast and sensitive searching and clustering of large protein and nucleotide sequence sets.

MMseqs2

License

Free to use and open source under GNU GPLv3.

Available

Roihu-CPU: 18-8cc5c, via the bio-apps module.
Roihu-GPU: 18-8cc5c (GPU-accelerated), via the bio-apps module.

On GPU nodes, the MMseqs2 build is compiled with CUDA support and can use the GH200 GPUs to accelerate searches.

Usage

MMseqs2 is part of the bio-apps collection on Roihu. Load the bio-apps module tree and then the MMseqs2 module:

module load bio-apps/v202603
module load mmseqs2/18-8cc5c

A typical search converts the query and target FASTA files to MMseqs2 databases and runs mmseqs easy-search:

mmseqs easy-search query.fasta target.fasta results.m8 tmp --threads 8

Example batch script

#!/bin/bash
#SBATCH --job-name=mmseqs2
#SBATCH --account=<project>
#SBATCH --output=output_%j.txt
#SBATCH --error=errors_%j.txt
#SBATCH --partition=small
#SBATCH --time=04:00:00
#SBATCH --nodes=1
#SBATCH --ntasks=1
#SBATCH --cpus-per-task=8
#SBATCH --mem-per-cpu=4G

module load bio-apps/v202603
module load mmseqs2/18-8cc5c

mmseqs easy-search query.fasta target.fasta results.m8 tmp --threads $SLURM_CPUS_PER_TASK

Replace <project> with your CSC project (for example project_2001234). To use GPU acceleration, run MMseqs2 in a GPU batch job on a GH200 node; see creating a batch job script for Roihu.

Support

CSC Service Desk