trash-py

Tandem-repeat array identifier — a Python port of the TRASH program.

Origin and acknowledgements

trash-py is a Python re-implementation of the TRASH program written by Piotr Włodzimierz (pwlodzimierz@ibb.waw.pl, Institute of Biochemistry and Biophysics, Polish Academy of Sciences). The upstream repository lives at https://github.com/vlothec/TRASH_2. All algorithmic credit for the underlying approach belongs to the original author; this port is an independent rewrite (no code is byte-identical with the upstream) that retains the structure and logic of the upstream pipeline.

The upstream MIT license is reproduced in LICENSES/TRASH-UPSTREAM-MIT.txt to satisfy its notice-preservation clause.

What's different?

trash-py aims to build on the substantial work done in developing the original TRASH repeat annotation pipeline by adopting a more flexible (and performant) Python/C runner/libs framework. A library of reusable functions is exposed which can be incorporated into diverse repeat-annotation related workflows beyond simply running the TRASH pipeline, and the hotter functions have been ported to C to maximise performance.

Currently, on smaller less repetitive genomes (e.g. Arabidopsis, Human genome) the bottleneck is nhmmer rather than the the TRASH pipeline itself, which takes a fraction of the original time to complete.

The stderr logs have also been changed to reflect the new internals and to give trash-py a little of its own personality.

Installation

conda

The best way to install the tool is using conda/mamba/micromamba:

conda install -c bioconda trash-py

source

If instead you'd like to build/install from source, please install nhmmer and Clustal Omega and ensure they are available on the PATH. Additionally, please ensure you have a suitable C compiler installed (gcc, clang) for the C extensions. There is no need to separately compile these, python should recognise the instructions for compilation in setup.py.

trash-py can then be installed by:

git clone https://github.com/mbeavitt/trash-py
cd trash-py
pip install .

Usage

trash-py -f input.fasta -o output_dir

Currently the CLI aims to mirror the one in the original TRASH tool as closely as possible, to present a drag-and-drop replacement. trash-py adds two options over upstream: -q to silence logs, and -p to run the array-identification and repeat-mapping stages across multiple worker processes.

Benchmarks

How to cite

If you use trash-py in academic work, please cite the original TRASH publication:

Wlodzimierz, P., Hong, M., & Henderson, I. R. (2023). TRASH: tandem repeat annotation and structural hierarchy. Bioinformatics, 39(5), btad308.

BibTeX:

@article{wlodzimierz2023trash,
  title={TRASH: tandem repeat annotation and structural hierarchy},
  author={Wlodzimierz, Piotr and Hong, Michael and Henderson, Ian R},
  journal={Bioinformatics},
  volume={39},
  number={5},
  pages={btad308},
  year={2023},
  publisher={Oxford University Press}
}

License

trash-py is released under the MIT License — see LICENSE.

Command-line reference

$ trash-py --help
usage: trash-py [-h] [-V] -f FASTA -o OUTPUT [-m MAX_REP_SIZE]
                [-i MIN_REP_SIZE] [-t TEMPLATES] [-q] [-p PROCESSES]

TRASH — tandem-repeat array identifier (Python)

options:
  -h, --help            show this help message and exit
  -V, --version         show program's version number and exit
  -f, --fasta FASTA     input fasta
  -o, --output OUTPUT   output directory
  -m, --max-rep-size MAX_REP_SIZE
  -i, --min-rep-size MIN_REP_SIZE
  -t, --templates TEMPLATES
                        optional template fasta — assigns class names from
                        headers
  -q, --quiet           suppress progress output
  -p, --processes PROCESSES
                        parallel worker processes for the array-identification
                        and repeat-mapping stages (default 1 = serial)