taf-subread

TAFFISH wrapper for Subread, the command-line suite that includes the Subread aligner, Subjunc, Sublong, featureCounts, ExactSNP, subindel, index building, and small utility programs for short-read mapping and read summarization workflows.

This app packages the official SourceForge Subread 2.1.1 source release. The GitHub repository currently has a stale latest tag/release at 2.0.2, while the GitHub master branch reports 2.0.6 from src/makefile.version. Upstream's current official SourceForge release and current Subread/Rsubread user guide are Subread v2.1.1, so this TAFFISH app uses 2.1.1.

Package metadata:

name: subread
command: taf-subread
version: 2.1.1-r2
kind: tool
image: ghcr.io/taffish/subread:2.1.1-r2
upstream release: SourceForge subread-2.1.1
runtime version: Subread-align v2.1.1 / featureCounts v2.1.1
native platforms: linux/amd64, linux/arm64

Install

After the TAFFISH Hub index has been updated:

taf update
taf install subread

Install the exact release:

taf install subread 2.1.1-r2

For local testing before the app is published to the public index:

taf install --from .

Usage

Show TAFFISH wrapper help and package version:

taf-subread --help
taf-subread --version

Show upstream Subread help and versions:

taf-subread -- -h
taf-subread -- -v
taf-subread featureCounts -v
taf-subread subread-buildindex -h

subread-align -v prints Subread-align v2.1.1 but exits non-zero in the upstream program. This is an upstream CLI quirk; the app smoke tests verify the printed version string with a shell pipeline.

The default upstream command is subread-align. Since TAFFISH command mode treats a non-option first argument as an executable inside the container, most suite commands should be called explicitly:

taf-subread subread-buildindex -o genome_index genome.fa
taf-subread subread-align -i genome_index -r reads.fq.gz -o aligned.bam
taf-subread subjunc -i genome_index -r rna_1.fq.gz -R rna_2.fq.gz -o subjunc.bam
taf-subread featureCounts -a genes.gtf -o counts.txt aligned.bam
taf-subread exactSNP -g genome.fa -i aligned.bam -b -o variants.vcf

Option-leading arguments can be passed to the default subread-align command with --:

taf-subread -- -v
taf-subread -- -i genome_index -r reads.fq.gz -o aligned.bam

Runtime Contents

The image builds Subread from the official subread-2.1.1-source.tar.gz source tarball and verifies its SHA256 during the image build:

6392d7c66831cdd767e58251892a79a51b6fab8ed0ba9671ad5e85ff1ab01eaa

Packaged top-level commands include:

• subread-align
• subread-buildindex
• subjunc
• sublong
• featureCounts
• exactSNP
• subindel

Packaged utility commands in /opt/subread/bin/utilities are also on PATH:

• detectionCall
• genRandomReads
• repair
• propmapped
• qualityScores
• removeDup
• subread-fullscan
• txUnique
• flattenGTF

The image keeps upstream documentation, annotation files, and test data under /opt/subread/doc, /opt/subread/annotation, and /opt/subread/test. The annotation directory includes upstream RefSeq exon annotations for human and mouse genomes such as hg19, hg38, mm9, mm10, and mm39.

Runtime dependencies are small: Debian glibc, zlib, bash/core utilities, gzip, and shared runtime libraries. Subread's SAM/BAM reading and writing code is built into the upstream C programs; samtools is not an external runtime dependency for the packaged command-line suite.

Boundaries

This app packages the SourceForge Subread command-line programs. It does not include the Bioconductor Rsubread R package, limma/voom, edgeR, or downstream RNA-seq differential-expression workflows. Use the command-line featureCounts included here for counting; use an R/Bioconductor environment separately for Rsubread-specific functions.

Genome indexes are not prebuilt. Build them with subread-buildindex from your own reference FASTA. Whole-genome indexes can require substantial memory and disk; smoke tests use a tiny bundled reference only.

The upstream test scripts are not used wholesale because some of them assume python2 and local relative paths. The app instead keeps the upstream test data and uses small direct smoke commands for featureCounts, index building, subread-align, Subjunc, and ExactSNP.

Platform

This app is declared native linux/amd64 and linux/arm64. The Dockerfile builds from source on Debian and removes an upstream -mtune=core2 tuning flag so the same source can compile natively on non-x86 targets without changing the reported Subread version or command behavior.

Licensing

Subread is distributed under GPLv3-or-later. The TAFFISH app repository LICENSE is Apache-2.0. The upstream LICENSE file remains included in the image at /opt/subread/LICENSE. The image also contains Debian runtime packages with their own license terms.

Smoke Coverage

Smoke tests are self-contained and run without network access. They check:

• Runtime version strings for subread-align, featureCounts, and subjunc.
• Help surfaces for subread-align, subread-buildindex, featureCounts, and exactSNP.
• Main and utility command availability.
• Dynamic library resolution for the compiled programs.
• A minimal featureCounts run on bundled SAM/GTF test data.
• A tiny subread-buildindex plus subread-align run.
• A tiny Subjunc paired-end SAM-output run.
• A minimal ExactSNP BAM-to-VCF run.

The smoke tests verify container integrity and representative basic behavior; they do not replace full alignment, counting, or variant-calling validation on real datasets.

License Boundary

The TAFFISH app packaging files are licensed under Apache-2.0. The packaged upstream Subread software is covered by: GPL-3.0-or-later. Bundled third-party components, datasets, models, and external resources keep their own license terms.

Upstream

• Source/homepage: https://subread.sourceforge.net/
• Official release files: https://sourceforge.net/projects/subread/files/subread-2.1.1/
• Source tarball: https://downloads.sourceforge.net/project/subread/subread-2.1.1/subread-2.1.1-source.tar.gz
• GitHub mirror: https://github.com/ShiLab-Bioinformatics/subread
• User guide: https://bioconductor.org/packages/devel/bioc/vignettes/Rsubread/inst/doc/SubreadUsersGuide.pdf

Recommended citations:

• Subread/Subjunc aligners: Liao, Smyth and Shi 2013, DOI 10.1093/nar/gkt214, PMID 23558742.
• featureCounts: Liao, Smyth and Shi 2014, DOI 10.1093/bioinformatics/btt656, PMID 24227677.
• Rsubread/Subread RNA-seq workflow: Liao, Smyth and Shi 2019, DOI 10.1093/nar/gkz114, PMID 30783653.