htseq-count-cluster

A cli wrapper for running htseq's htseq-count on a cluster.

View documentation.

Install

Requires Python 3.9 or higher.

pip install HTSeqCountCluster

Features

• For use with large datasets (we've previously used a dataset of 120 different human samples)
• For use with SGE/SGI cluster systems
• Submits multiple jobs
• Command line interface/script
• Merges counts files into one counts table/csv file
• Uses accepted_hits.bam file output of tophat

Examples

Run htseq-count-cluster

After generating bam output files from tophat, instead of using HTSeq's htseq-count, you can use our htseq-count-cluster script. This script is intended for use with clusters that are using pbs (qsub) for job monitoring.

Our default htseq-count command is htseq-count -f bam -s no file.bam file.gtf -o htseq.out. This command does not take into account any strandedness (-s no) for the input bamfiles (-f bam) and uses the default union mode. For the default mode union, only the aligned read determines how the read pair is counted.

Legacy mode (still supported):

htseq-count-cluster -p path/to/bam-files/ -f samples.csv -g genes.gtf -o path/to/cluster-output/

New subcommand mode:

htseq-count-cluster run -p path/to/bam-files/ -f samples.csv -g genes.gtf -o path/to/cluster-output/

This script uses logzero so there will be color coded logging information to your shell.

A common linux practice is to use screen to create a new shell and run a program so that if it does produce output to the stdout/shell, the user can exit that particular shell without the program ending and utilize another shell.

Help message output for htseq-count-cluster

usage: htseq-count-cluster [-h] COMMAND ...

This is a command line wrapper around htseq-count.

positional arguments:
  COMMAND
    run                 Run htseq-count jobs on a cluster
    merge               Merge multiple counts tables into one CSV file

optional arguments:
  -h, --help            show this help message and exit

*Ensure that htseq-count is in your path.

For the run subcommand:

usage: htseq-count-cluster run [-h] -p INPATH -f INFILE -g GTF -o OUTPATH [-e EMAIL]

Submit multiple htseq-count jobs to a cluster.

optional arguments:
  -h, --help            show this help message and exit
  -p INPATH, --inpath INPATH
                        Path of your samples/sample folders.
  -f INFILE, --infile INFILE
                        Name or path to your input csv file.
  -g GTF, --gtf GTF     Name or path to your gtf/gff file.
  -o OUTPATH, --outpath OUTPATH
                        Directory of your output counts file. The counts file
                        will be named.
  -e EMAIL, --email EMAIL
                        Email address to send script completion to.

Merge output counts files

In order to prep your data for DESeq2, limma or edgeR, it's best to have 1 merged counts file instead of multiple files produced from the htseq-count-cluster script.

Using the merge subcommand:

htseq-count-cluster merge -d path/to/cluster-output/

Or using the standalone command (still available):

merge-counts -d path/to/cluster-output/

Help message for merge subcommand

usage: htseq-count-cluster merge [-h] -d DIRECTORY

Merge multiple counts tables into 1 counts .csv file.

Your output file will be named:  merged_counts_table.csv

optional arguments:
  -h, --help            show this help message and exit
  -d DIRECTORY, --directory DIRECTORY
                        Path to folder of counts files.

ToDo

• Monitor jobs.
• Enhance wrapper input for other use cases.
• Add example data.

Maintainers

Shaurita Hutchins | @sdhutchins | ✉
Rob Gilmore | @grabear | ✉

Help

Please feel free to open an issue if you have a question/feedback/problem or submit a pull request to add a feature/refactor/etc. to this project.

Citation

Simon Anders, Paul Theodor Pyl, Wolfgang Huber; HTSeq—a Python framework to work with high-throughput sequencing data, Bioinformatics, Volume 31, Issue 2, 15 January 2015, Pages 166–169, https://doi.org/10.1093/bioinformatics/btu638