MaxQuant with snakemake and singularity workflow for open and scalable mass spectrometry data analysis on Linux computing clusters
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
MaxQuant
README.md
UltraQuant.sm
config.yaml

README.md

UltraQuant

Introduction

UltraQuant is an open framework for running MaxQuant on Linux computer clusters written in the Snakemake workflow management system.

Installation

Dependencies

UltraQuant uses Miniconda to install the Snakemake workflow management system. The following dependencies are either included as part of UltraQuant or automatically installed using Singularity container to be used in diverse Linux environments:

Installing UltraQuant

UltraQuant can be installed using git clone:

git clone https://github.com/kentsisresearchgroup/UltraQuant

Configuration

To run UltraQuant, create a copy of the user configuration "config.yaml" file provided and adjust the locations of files and directories.

Directories

  • "work_directory" sets the path that will serve as the relative path for all analyses. The user should create a new directory for each dataset that contains mass spectrometry data files. In a cluster environment, this path should be accessible to all nodes.
  • "temporary_directory" specifies the directory which MaxQuant uses for writing and reading temporary files. This can be a local scratch storage directory that is mounted locally on each node on a cluster. Alternatively this can be a directory on a high-performance network storage file system, if network access is sufficiently fast.

MaxQuant

MaxQuant requires mass spectrometry data as RAW files, search database as FASTA file, and a parameter file.

  • "database_file" specifies the user-provided search database FASTA file.
  • "raw_directory" specifies the directory containing all user-provided .raw mass spectrometry data files. UltraQuant will run MaxQuant on all .raw files in this directory. To perform calculations on independent raw files, they should be organized in individual directories.
  • "threads" specifies the number of computational threads used by MaxQuant.

The variables below are specified in the Snakefile "UltraQuant.sm". These files are provided and should not be changed.

  • "MQ" specifies the MaxQuant executable file, which is provided in "/MaxQuant/bin/MaxQuantCmd.exe" as part of this container distribution. The currently provided version of MaxQuant is 1.6.2.3, including its license agreement for redistribution located in "/MaxQuant". To switch to alternate versions of MaxQuant, please update both the executable and the parameter file.
  • "PAR" specifies the template MaxQuant parameter file, which is provided in "/MaxQuant" and is used to generate specific search parameters, including file names, locations, as well as mass spectrum processing and analysis parameters.

To alter the mass spectrum processing and analysis parameters, you should create a new MaxQuant parameter file by executing "/MaxQuant/bin/MaxQuantGui.exe" using Windows operating system. Click File > Load parameters... and load "/MaxQuant/mqpar_template.xml" and adjust the parameters. DO NOT add or remove RAW and FASTA files as they are specified by UltraQuant. Finally, click File > Save parameters... and save it as "mqpar.xml" in the same directory as the template MaxQuant parameter file. UltraQuant will automatically detect this file and use it as input for MaxQuant instead of its self-generated parameters. Alternatively, you may edit the "PAR" variable to specify alternative search parameter xml files.

Running UltraQuant

This example demonstrates how to run UltraQuant on an LSF cluster. You can parallelize multiple jobs by specifying separate "--configfile" arguments.

snakemake --snakefile UltraQuant.sm --configfile config.yaml --cluster \
"bsub -J 'myjob.{params.J}' -n {params.n} -R {params.R} -W 16:00 -o 'myjob.{params.o}' -eo 'myjob.{params.eo}'" \
--jn {rulename}.{jobid}.sj -j 50 -k --latency-wait 60 --use-singularity --singularity-args \
"--bind /data:/data,/lila:/lila,/scratch:/scratch" --ri
  • "-J" specifies the job name.
  • "-n" specifies the number of requested calculation threads.
  • "-R" specifies the resources necessary for the the job. "span[host=1]" specifies the number of nodes used for parallelization, if implemented in specific cluster environments. "rusage[mem=15]" specifies the minimum memory in GB requested per node. "span[ptile=32]" specifies the number of requested processors used per node. MaxQuant runs faster when hyperthreaded on the same node. These values can be adjusted to fit specific cluster configurations.
  • "-W" specific the maximum time allowed for the job in hours.
  • "-o" specifies the output file name.
  • "-eo" specifies the error file name.

"--bind" implements binding of directories for UltraQuant to access files outside of this Singularity container, including the directories for UltraQuant itself, "TMP" calculation directory, and "RAW" mass spectrometry files directory.

Expected Output

UltraQuant will execute MaxQuant, which will generate result tables in the user-specified working directory, under "out/combined/txt/".

Authors