srf

TAFFISH wrapper for SRF, Heng Li's Satellite Repeat Finder for assembling satellite repeat units from high-occurrence k-mer count tables.

Upstream SRF has no semantic release tag or runtime version flag. This app therefore packages the upstream HEAD commit e54ca8c8eccf6b1f19428b0f862f2c90575290a0 as srf 20240109-r2. The version is the commit date in Asia/Shanghai time; the same commit is 2024-01-08T21:02:55Z in UTC.

Package metadata:

name: srf
command: taf-srf
version: 20240109-r2
kind: tool
image: ghcr.io/taffish/srf:20240109-r2
upstream commit: e54ca8c8eccf6b1f19428b0f862f2c90575290a0

Release 20240109-r2 is a help-only/docs-only correction for the TAFFISH wrapper help. It keeps the same upstream SRF commit, Dockerfile, runtime dependencies, smoke coverage, and command behavior as 20240109-r1.

Quick Start

Show the TAFFISH package version:

taf-srf --version

Show upstream SRF usage:

taf-srf -- --help
taf-srf srf

Assemble satellite repeat contigs from a KMC dump-style k-mer count table:

taf-srf srf -p sample count.txt > sample.srf.fa

Because srf is the default upstream command, option-leading SRF calls can also use the short form:

taf-srf -p sample -l 100 count.txt > sample.srf.fa

Run helper commands through command mode:

taf-srf srfutils.js enlong sample.srf.fa > sample.srf.enlong.fa
taf-srf srfutils.js paf2bed sample.srf.paf > sample.srf.bed
taf-srf srfutils.js bed2abun sample.srf.bed > sample.srf.abun.txt

Input Contract

The primary srf executable reads a plain text or gzip-compressed k-mer count table. Each line should contain one k-mer and one positive count:

ACGTTGCAACGTTGCAA 125
CGTTGCAACGTTGCAAC 123

Upstream usually expects this table to come from kmc_dump after high-copy k-mers have been selected with KMC. The exact -k, -ci, and -cs choices depend on the input data and coverage; SRF itself consumes the dumped table, not the original reads or contigs.

Pass input files from the current project directory or another path mounted by TAFFISH, such as your home directory. Upstream SRF does not always produce a friendly diagnostic when a count file is missing or not visible inside the container.

Helper Tools

The image contains:

srf
srfutils.js
fmd-occ
k8
gzip

srfutils.js is upstream's JavaScript helper for downstream analysis. It is installed as an executable script, and the image includes the k8 JavaScript engine needed by its shebang. Common helper commands are:

enlong    elongate circular contigs before mapping
filter    filter read-to-contig PAF alignments
paf2bed   extract non-overlapping satellite regions
bed2cnt   summarize per-chromosome counts
bed2abun  estimate per-contig abundance
merge     merge identical sequences

The Makefile also defines fmd-occ; it is installed for completeness as an upstream low-level helper, but the public README workflow centers on srf and srfutils.js.

Common Workflow

SRF does not perform k-mer counting by itself. A typical upstream workflow is:

kmc -fq -k151 -t16 -ci100 -cs100000 reads.fq.gz count.kmc tmp_dir
kmc_dump count.kmc count.txt
taf-srf srf -p sample count.txt > sample.srf.fa

For abundance analysis, map the elongated SRF contigs back to reads or an assembly with an aligner such as minimap2, then process the PAF:

taf-srf srfutils.js enlong sample.srf.fa > sample.srf.enlong.fa
minimap2 -c -N1000000 -f1000 -r100,100 sample.srf.enlong.fa reads.fa > sample.srf.paf
taf-srf srfutils.js paf2bed sample.srf.paf > sample.srf.bed
taf-srf srfutils.js bed2abun sample.srf.bed > sample.srf.abun.txt

Container Contents

The Dockerfile builds SRF from the pinned upstream commit on debian:12-slim. It compiles the C sources with GCC, GNU Make, and zlib headers, installs the compiled srf and fmd-occ binaries, installs srfutils.js, and adds Bioconda-packaged k8 1.2 binaries pinned by SHA256 for both supported architectures.

Runtime dependencies are intentionally small:

bash
gzip
libgcc-s1
zlib1g

Source metadata is installed at:

/opt/srf/share/doc/srf/source.txt

The image is built and validated for:

linux/amd64
linux/arm64

Boundaries

This app packages SRF and its upstream helper script. It does not bundle KMC, minimap2, TRF, TRF-mod, read mappers, plotting tools, reference genomes, read sets, or satellite-repeat databases. Use separate TAFFISH tools or host tools for those workflow stages.

Upstream has no --version command. Version reproducibility is provided by the TAFFISH package version and by the pinned commit/source SHA256 recorded in the image.

Input path handling follows normal TAFFISH container mounts. Files outside the mounted working tree or home directory need to be copied or bound into a mounted location before they are passed to srf.

Build Notes

The Dockerfile downloads:

https://github.com/lh3/srf/archive/e54ca8c8eccf6b1f19428b0f862f2c90575290a0.tar.gz
sha256: 8a9e6851ff02c119eade2b1a8f1758ba0fda4780cd8547630da204abf84ba6c4

The included k8 runtime is the same Bioconda k8 1.2 package family used by the minimap2 app in this hub, with architecture-specific SHA256 pins.

Smoke Coverage

The package smoke checks are self-contained and offline. They verify:

srf, srfutils.js, fmd-occ, k8, gzip, ldd, and sh command existence
pinned source metadata for version 20240109 and commit e54ca8c8...
upstream usage text for srf, srfutils.js, and fmd-occ
zlib linkage for srf and k8
a tiny k-mer count table producing an SRF FASTA record
gzip-compressed k-mer count input
srfutils.js enlong, merge, and bed2abun helper paths

These checks validate packaging, command surface, and minimal local execution; they do not replace scientific validation of SRF parameters on real sequencing data.

Upstream

• Project: SRF, Satellite Repeat Finder
• Source: https://github.com/lh3/srf
• Pinned commit: https://github.com/lh3/srf/commit/e54ca8c8eccf6b1f19428b0f862f2c90575290a0
• Upstream license: MIT
• Citation: Zhang Y, Chu J, Cheng H and Li H (2023), arXiv:2304.09729

Maintainer Notes

Useful checks before publishing:

taf check
taf compile -- -p sample count.txt
taf publish --release --dry-run
docker build -t ghcr.io/taffish/srf:20240109-r2 -f docker/Dockerfile .
docker run --rm ghcr.io/taffish/srf:20240109-r2 srf
docker run --rm ghcr.io/taffish/srf:20240109-r2 srfutils.js