repeatmasker

TAFFISH app for RepeatMasker, a widely used tool for screening DNA sequences for interspersed repeats and low-complexity DNA.

Package Identity

• name: repeatmasker
• command: taf-repeatmasker
• version: 4.2.4-r1
• kind: tool
• image: ghcr.io/taffish/repeatmasker:4.2.4-r1
• upstream: RepeatMasker 4.2.4
• runtime version: RepeatMasker version 4.2.4
• default command: RepeatMasker through taf-repeatmasker-entrypoint
• default search engine: rmblast
• native platform: linux/amd64

The image is configured non-interactively with RMBlast, TRF, HMMER, and small UCSC helper tools used by RepeatMasker utilities. It also packages FamDB 3.0.0. PAGER=cat is set so upstream RepeatMasker help prints and exits instead of entering more. The app is intended for offline custom-library runs and for Dfam/FamDB runs when users provide the required Dfam component files.

Install

taf install repeatmasker

Usage

TAFFISH wrapper help:

taf-repeatmasker --help
taf-repeatmasker --version
taf-repeatmasker --compile

Upstream RepeatMasker help and version:

taf-repeatmasker -- -help
taf-repeatmasker -- --help
taf-repeatmasker -- -v
taf-repeatmasker RepeatMasker -v

Run with a custom repeat library:

taf-repeatmasker -pa 8 -engine rmblast -lib repeats.fa -dir rm-out genome.fa

Run bundled utility commands through command mode:

taf-repeatmasker famdb.py -i /opt/RepeatMasker/Libraries/famdb info
taf-repeatmasker rmOutToGFF3.pl genome.fa.out > genome.fa.out.gff
taf-repeatmasker buildSummary.pl genome.fa.out > genome.fa.tbl
taf-repeatmasker rmToTrackHub.pl -out genome.fa.out -genome hg38
taf-repeatmasker rmblastn -version
taf-repeatmasker trf -v
taf-repeatmasker famdb.py --help
taf-repeatmasker download_dfam.py --help

Wrapper Semantics

taf-repeatmasker defaults to a small TAFFISH entrypoint that runs upstream RepeatMasker. Because RepeatMasker uses single-dash options such as -pa, -lib, -species, and -dir, normal option-led calls are passed directly to that default entrypoint.

RepeatMasker 4.2.4 fails early if FamDB is configured but no Dfam .h5 files are present. To keep custom-library -lib runs working without Dfam data, the image leaves FamDB out of the static RepeatMasker config. The entrypoint auto-enables FamDB only when mounted .h5 component files are found at runtime.

Command mode remains enabled. If the first argument is another executable name, TAFFISH runs that command in the same configured container:

taf-repeatmasker famdb.py --help
taf-repeatmasker ProcessRepeats -help
taf-repeatmasker rmblastn -version

Runtime Contents

The image installs the official RepeatMasker-4.2.4.tar.gz release, installs FamDB 3.0.0, copies RMBlast from the ghcr.io/taffish/rmblast:2.17.1-r1 app image, and installs pinned UCSC command-line helper binaries needed by RepeatMasker utility scripts. The Dockerfile verifies the RepeatMasker, FamDB, and UCSC helper checksums:

RepeatMasker-4.2.4.tar.gz
7407b42909c64b7ae080efb737f3f725327be81ace1ec0cd9dc13c4a503885d7

FamDB-3.0.0.tar.gz
cad4652696cbd6b3444ef1ea008764e4807eba503410bf24357c4505a5ddb8d4

twoBitInfo
fca5427e1aedd00be0d32117420010c9162d442562c83305b2d3e3c8e6cebe71

bedToBigBed
73e10feef3a685304d71c840058ae9a5113bdc4e134f161d3913cfebc5a2ec24

The RMBlast source checksum is enforced by the separate rmblast taf-app:

rmblast-2.17.1+-x64-linux.tar.gz
9ddf0ffdc01f3dd7bf62bbcb01516757b1e732e54674e81d93a7c143157aa527

Packaged RepeatMasker commands and utilities include:

• RepeatMasker
• taf-repeatmasker-entrypoint
• ProcessRepeats
• RepeatProteinMask
• DupMasker
• famdb.py
• download_dfam.py
• RM2Bed.py
• rmOutToGFF3.pl
• buildSummary.pl
• calcDivergenceFromAlign.pl
• createRepeatLandscape.pl
• rmToTrackHub.pl
• maskFile.pl
• buildRMLibFromEMBL.pl

Search/runtime tools include:

• RMBlast 2.17.1+ (rmblastn, makeblastdb, and companion BLAST+ tools)
• TRF 4.09.1 from Debian
• HMMER 3.3.2 from Debian
• FamDB 3.0.0 with Python h5py/numpy support
• UCSC twoBitInfo and bedToBigBed helper binaries
• Perl, Python 3, Python h5py, curl, gzip, tar, and required shared libraries

BLAST_USAGE_REPORT=false is set by default so RMBlast does not report usage statistics during local TAFFISH runs. PAGER=cat is set by default so RepeatMasker help remains non-interactive and works predictably in terminals, pipelines, CI, and TAFFISH flow contexts.

Libraries And Data

RepeatMasker 4.2.4 no longer bundles a local Dfam root partition. Upstream now treats FamDB as a separate optional package and Dfam 4.0 as external component data. This image includes the FamDB 3.0.0 program and configures FamDB itself to look for data in:

/opt/RepeatMasker/Libraries/famdb

Custom-library runs with -lib do not need FamDB data. Species or clade queries with -species require Dfam component files. When .h5 component files are mounted at the path above, the default TAFFISH entrypoint configures RepeatMasker with /opt/FamDB before running the command. For most RepeatMasker use cases, upstream FamDB recommends downloading the root component plus curated consensus files; curated HMM files add sensitivity. A complete Dfam 4.0 download is very large and is not bundled.

The wrapper automatically mounts an existing local Dfam/FamDB directory to /opt/RepeatMasker/Libraries/famdb when one of these paths exists, and the entrypoint then enables FamDB for RepeatMasker:

TAFFISH_REPEATMASKER_FAMDB_PATH
~/.local/share/taffish/databases/repeatmasker/dfam4
/usr/local/share/taffish/databases/repeatmasker/dfam4
/opt/taffish/databases/repeatmasker/dfam4

Personal setup example:

mkdir -p ~/.local/share/taffish/databases/repeatmasker/dfam4
cd ~/.local/share/taffish/databases/repeatmasker
taf-repeatmasker download_dfam.py -o "$PWD/dfam4"
taf-repeatmasker famdb.py -i "$PWD/dfam4" info

Administrator setup can use /usr/local/share/taffish/databases/repeatmasker with the same dfam4 directory name, then make the files readable by users. For one-off data elsewhere, set:

TAFFISH_REPEATMASKER_FAMDB_PATH=/abs/path/dfam4 \
taf-repeatmasker -pa 8 -engine rmblast -species "Homo sapiens" genome.fa

To disable auto-mount and provide backend args manually:

TAFFISH_REPEATMASKER_AUTO_MOUNT=0 \
TAFFISH_DOCKER_RUN_ARGS="-v /abs/path/dfam4:/opt/RepeatMasker/Libraries/famdb:ro" \
taf-repeatmasker -pa 8 -engine rmblast -species "Homo sapiens" genome.fa

Use TAFFISH_PODMAN_RUN_ARGS for Podman. Symlinks are resolved to their physical target before auto-mounting, so the target must be readable and visible to the selected container backend.

RepBase is not included. RepBase RepeatMasker Edition and later RepBase data must be obtained by users under their own authorization and used according to their license terms.

Platform

This app is native linux/amd64 only because it uses the official RMBlast x64 Linux binary. src/main.taf asks Docker and Podman to run with --platform linux/amd64. On arm64 hosts this is amd64 emulation, not native arm64 support. Apptainer behavior depends on the host's ability to run amd64 containers.

Boundaries

This core image does not include:

• full Dfam component data
• RepBase
• cross_match
• ABBLAST/WUBlast
• a from-source RMBlast build inside this Dockerfile; it depends on the ghcr.io/taffish/rmblast:2.17.1-r1 app image at build time
• Entrez Direct. The upstream download_dfam.py utility is packaged for explicit user-initiated Dfam downloads, and utility scripts such as rmToTrackHub.pl can still use curl to fetch UCSC chromosome sizes when users request those workflows.

The smoke tests validate the configured program stack and small offline functional runs. They do not download Dfam data and do not substitute for full biological validation on large genomes or species-specific repeat libraries.

License Boundary

The TAFFISH app packaging files are licensed under Apache-2.0. The packaged upstream RepeatMasker software is covered by OSL-2.1. Packaged FamDB code and Dfam data notices are CC0, while RMBlast/NCBI components and UCSC command-line utilities keep their own terms. External Dfam component files, RepBase data, models, and other user-provided resources keep their own license terms.

License And Citation

RepeatMasker open-4.0 and later are distributed under the Open Software License v2.1. The packaged FamDB code is documented by upstream as CC0. RMBlast and NCBI BLAST+ notices are included through the RMBlast distribution. UCSC twoBitInfo and bedToBigBed are included as command-line utilities for RepeatMasker helper-script compatibility; consult the UCSC Genome Browser/Kent source terms for those tools.

Please refer to:

• Smit, AFA, Hubley, R. & Green, P. RepeatMasker at https://www.repeatmasker.org/
• Dfam Consortium for Dfam/FamDB data
• Camacho et al. 2009 for BLAST+. DOI: 10.1186/1471-2105-10-421; PMID: 20003500

Smoke Coverage

Smoke tests cover:

• RepeatMasker, RMBlast, TRF, HMMER, and usage-reporting version/state checks.
• RepeatMasker non-interactive --help and ProcessRepeats help.
• FamDB 3.0.0 help, Dfam downloader help, and missing-Dfam-data boundary.
• RepeatMasker configuration paths for RMBlast, TRF, HMMER, and default engine.
• A tiny offline default-entrypoint -lib run producing .out and .masked.
• A tiny offline default-entrypoint -gff -lib run producing .out.gff.
• Utility conversion/summary scripts on a tiny real RepeatMasker .out file, including the UCSC twoBitInfo helper required by buildSummary.pl.