Skip to content

ComparativeGenomicsToolkit/cactus

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Cactus

Build Status

Cactus is a reference-free whole-genome alignment program, as well as a pangenome graph construction toolkit.

Getting Cactus

  • Use the precompiled binaries (Linux X86) or Docker image from the latest release
  • See below for details on building from source.

Getting help

Please subscribe to the cactus-announce low-volume mailing list so we may reach out about releases and other announcements.

To ask questions or request help, please use the Cactus GitHub Discussions.

To file a bug report or enhancement request against the code or documentation, create a GitHub Issue.

Align Genomes from Different Species

Align Genomes from the Same Species and Build Pangenome Graphs

Acknowledgements

Cactus uses many different algorithms and individual code contributions, principally from Joel Armstrong, Glenn Hickey, Mark Diekhans and Benedict Paten. We are particularly grateful to:

  • Yung H. Tsin and Nima Norouzi for contributing their 3-edge connected components program code, which is crucial in constructing the cactus graph structure, see: Tsin,Y.H., "A simple 3-edge-connected component algorithm," Theory of Computing Systems, vol.40, No.2, 2007, pp.125-142.
  • Bob Harris for providing endless support for his LastZ pairwise, blast-like genome alignment tool.
  • Melissa Jane Hubiz and Adam Siepel for halPhyloP and Phast.
  • B Gulhan, R Burhans, R Harris, M Kandemir, M Haeussler, A Nekrutenko for KegAlign, the GPU-accelerated version of LastZ.
  • Yan Gao et al. for abPOA
  • Heng Li for minigraph, minimap2, gfatools and dna-brnn
  • Dany Doerr for GFAffix, used to optionally clean pangenome graphs.
  • The vg team for vg, used to process pangenome graphs.
  • The authors of Mash
  • Andrea Guarracino, Erik Garrison and co-authors for odgi. Make sure to cite odgi when using it or its visualizations.
  • Hani Z. Girgis for RED
  • Erik Garrison and co-authors for vcfwave. vcflib citation

Installing Manually From Source

Cactus requires Python >= 3.8 along with Python development headers and libraries

Clone cactus and submodules

git clone https://github.com/ComparativeGenomicsToolkit/cactus.git --recursive

Create the Python virtual environment. Install virtualenv first if needed with python3 -m pip install virtualenv.

cd cactus
virtualenv -p python3 cactus_env
echo "export PATH=$(pwd)/bin:\$PATH" >> cactus_env/bin/activate
echo "export PYTHONPATH=$(pwd)/lib:\$PYTHONPATH" >> cactus_env/bin/activate
source cactus_env/bin/activate
python3 -m pip install -U setuptools pip wheel
python3 -m pip install -U .
python3 -m pip install -U -r ./toil-requirement.txt

If you have Docker installed, you can now run Cactus. All binaries, such as lastz and cactus-consolidated will be run via Docker. Singularity binaries can be used in place of docker binaries with the --binariesMode singularity flag. Note, you must use Singularity 2.3 - 2.6 or Singularity 3.1.0+. Singularity 3 versions below 3.1.0 are incompatible with cactus (see issue #55 and issue #60).

By default, cactus will use the image corresponding to the latest release when running docker binaries. This is usually okay, but can be overridden with the CACTUS_DOCKER_ORG and CACTUS_DOCKER_TAG environment variables. For example, to use GPU release 2.4.4, run export CACTUS_DOCKER_TAG=v2.4.4-gpu before running cactus.

Compiling Binaries Locally

In order to compile the binaries locally and not use a Docker image, you need some dependencies installed. On Ubuntu (we've tested on 20.04 and 22.04), you can look at the Cactus Dockerfile for guidance. To obtain the apt-get command:

grep apt-get Dockerfile | head -1 | sed -e 's/RUN //g' -e 's/apt-get/sudo apt-get/g'

Progressive Cactus can be built on ARM cpus including on Mac (with packages installed via Brew), but Minigraph-Cactus is currently X86-only.

To build Cactus, run

make -j 8

In order to run the Minigraph-Cactus pipeline, you must also run

build-tools/downloadPangenomeTools

In order to run cactus-pangenome --vcfwave you may need to then run

export LD_LIBRARY_PATH=$(pwd)/lib:$LD_LIBRARY_PATH

If you want to work with MAF, including running cactus-hal2maf, you must also run

build-tools/downloadMafTools

In order to toggle between local and Docker binaries, use the --binariesMode command line option. If --binariesMode is not specified, local binaries will be used if found in PATH, otherwise a Docker image will be used.