avianalter/comparative_annotation_toolkit_hack repository overview

Comparative Annotation Toolkit is the recommended software for using progressivecactus to lift over annotations between related species, incorporating short- and long-read RNA-Seq data if available. The software has basically been orphaned since 2020, and neither the biocontainer or quay.io versions of CAT actually function properly, so I attempted to compile from the Dockerfile on github in 02/2023. I spent entirely too much time hacking at it to get it to work, and there are some closed issues on the CAT github roughly documenting some of the hacks I had to use to get the software to work. This image should have a "working" version of CAT that successfully completes on the provided test data in local binary mode only. It also doesn't provide completion statistics or the plots that are supposed to be generated. It may also say that the run did not complete correctly, but the lifted over annotations are available and complete.

Figuring out the rest of those issues were not really worth my time, as it turns out the software dies on my real data. Perhaps there may be a small fix that could've solved the issue, but I decided instead to try to use both HALPER and LiftOff. If someone builds a container/image based on this that actually gets CAT to work properly on large data sets or actually have used this software successfully, do not hesitate to contact me, as I would love to get it to work on my data. Beyond that, I cannot provide any support on this docker image.

Code for getting CAT running on test data below. Note, the only difference from the command on the CAT man page is the addition of the '--binary-mode local' flag:

$> cd /Comparative-Annotation-Toolkit

$> luigi --module cat RunCat --hal=test_data/vertebrates.hal --ref-genome=mm10 --workers=10 --config=test_data/test.config \ --work-dir test_install --out-dir test_install --local-scheduler --binary-mode local --augustus --augustus-cgp --augustus-pb --assembly-hub > log.txt