clipkit

clipkit packages ClipKIT for TAFFISH. ClipKIT trims multiple sequence alignments for phylogenetic and phylogenomic analysis while preserving phylogenetically informative sites.

Package identity:

• name: clipkit
• command: taf-clipkit
• kind: tool
• version: 2.12.2-r1
• image: ghcr.io/taffish/clipkit:2.12.2-r1
• upstream: JLSteenwyk/ClipKIT tag v2.12.2
• upstream commit: fa656cb6532aa3fb4c22ecef3b44721c658e283d
• runtime version: clipkit 2.12.2
• TAFFISH app license: Apache-2.0
• upstream license: MIT

What This App Packages

The image installs the official PyPI clipkit==2.12.2 wheel into a Python 3.11 virtual environment. The official PyPI source distribution is retained for README, package metadata, and license provenance. Both artifacts are pinned by SHA-256:

Artifact	SHA-256
clipkit-2.12.2-py3-none-any.whl	b9074ba75ef0a1217b2c4c7aed36ca927c3664117d59f28a9844cc539c6c0630
clipkit-2.12.2.tar.gz	bdc1e7f0d8da34e8a8622715de7203edf51da47d8163cc78799b923a7be0e2b4
upstream test.ecomp fixture	f9f9b2a1084757586a797708a5198156347164516393462a04a5cd624d568065

GitHub release assets and PyPI artifacts are separate builds with different digests. This app intentionally uses the PyPI URLs and PyPI-published digests listed above; checksums from the GitHub release page must not be substituted.

Runtime packages remain pinned to the validated Python 3.11 combination:

• clipkit==2.12.2
• biopython==1.87
• numpy==1.26.4
• zstandard==0.25.0

zstandard is included because zstd-encoded Evolutionary Compression (.ecomp) archives require the optional module. Other .ecomp encodings are handled by ClipKIT itself.

Upstream 2.12.2 Changes

Compared with 2.12.0, upstream now constructs NumPy alignment matrices directly from complete sequence strings, counts character states in compact column batches, reuses cached gap statistics for smart-gap selection, and materializes output rows directly as strings. Upstream reports identical trimming decisions with substantially lower runtime and memory on medium and large alignments.

The public CLI, supported formats, trimming modes, and Python dependency bounds did not add a new runtime dependency in this update.

Scope

ClipKIT supports smart-gap, gap-fraction, entropy, composition-bias, heterotachy, parsimony-informative, custom-site, codon-position, and related trimming modes. It can emit logs, complementary alignments, JSON summaries, and standalone HTML trim reports.

This app does not include an aligner, phylogenetic tree inference software, model selection, downstream workflows, or reference datasets. Input must already be a multiple sequence alignment.

Usage

Run ClipKIT explicitly through TAFFISH command mode:

taf-clipkit clipkit alignment.fa -o alignment.clipkit.fa
taf-clipkit clipkit alignment.fa -m smart-gap -o smart-gap.fa
taf-clipkit clipkit alignment.fa -m gappy -g 0.8 -o gappy.fa
taf-clipkit clipkit alignment.fa -m kpic -o kpic.fa
taf-clipkit clipkit alignment.fa -m c3 -co -o no-third-codon.fa
taf-clipkit clipkit archive.ecomp -of fasta -o trimmed.fa

Generate provenance and report outputs:

taf-clipkit clipkit alignment.fa -l -c \
  --report_json report.json \
  --plot_trim_report report.html \
  -o trimmed.fa

Access wrapper and upstream help:

taf-clipkit --help
taf-clipkit -- --help
taf-clipkit clipkit --help
taf-clipkit -- --version

Command Mode

The upstream CLI requires its input alignment as the first positional argument. Because TAFFISH automatic command mode treats a first non-option argument as an executable, use:

taf-clipkit clipkit alignment.fa -o trimmed.fa

Do not use taf-clipkit alignment.fa ...; that would try to execute a program named alignment.fa. Explicit command mode also exposes the packaged Python interpreter for script files:

taf-clipkit python analysis.py

Inputs and Outputs

ClipKIT accepts FASTA, Clustal, MAF, Mauve, PHYLIP, Stockholm, and .ecomp alignments through its upstream parsers. Output defaults to the input format; use -of fasta or another supported format to override it.

Depending on options, ClipKIT writes the trimmed alignment, a per-site log, the removed-site complementary alignment, JSON run metadata, and an HTML trim report. Outputs are ordinary user files in the selected working directory.

Platform and Resources

The app targets native linux/amd64 and linux/arm64. The Python package is architecture-neutral; NumPy, Biopython, and zstandard provide compiled wheels for both declared platforms. Memory use scales with alignment cells. ClipKIT 2.12.2 specifically reduces construction and counting overhead for larger MSAs.

No network, database, model, or external reference is needed at runtime.

Testing

Independent offline smoke covers:

• exact ClipKIT version, Python imports, source provenance, and native libraries
• real FASTA trimming with log, complement, JSON, and HTML report outputs
• validate-only mode and .ecomp decoding to FASTA
• a deterministic matrix above one million alignment cells that exercises the new batched count path and checks gap, entropy, composition, and frequency results

Dockerfile build-time checks intentionally use only a tiny alignment and normal CLI markers. The larger optimization-path test and HTML report generation stay in runtime smoke so multi-architecture buildx/QEMU builds are not burdened by unnecessarily strong build-time tests.

Smoke validates packaging and representative execution, not every mode or biological interpretation on production alignments.

License and Citation

The TAFFISH app packaging is Apache-2.0. ClipKIT retains its MIT license, whose text is included in the image.

Please cite Steenwyk et al., ClipKIT: a multiple sequence alignment trimming software for accurate phylogenomic inference, PLOS Biology (2020), DOI 10.1371/journal.pbio.3001007.

Upstream resources:

• https://github.com/JLSteenwyk/ClipKIT
• https://github.com/JLSteenwyk/ClipKIT/releases/tag/v2.12.2
• https://jlsteenwyk.com/ClipKIT/