-
Notifications
You must be signed in to change notification settings - Fork 172
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
kallisto|bustools 0.28.2 update #294
Conversation
|
Patch release v2.5.1
I am working on testing the current version to have a baseline. There is a potential existing bug: using the scrnaseq/workflows/scrnaseq.nf Line 147 in 90cb6a4
I suspect that this is because
But I do not believe these variables ( t1c and t2c ) are ever defined if the index is provided:
I assume fixing this bug would require changing the
|
That is not really what this argument does – it just concatenates the gene_version and transcript_version fields to IDs.
The txp2gene file should really be generated by kallisto|bustools. The t2g.py script generates a 3-column rather than 7-column t2g, which seems to be incompatible with kb count.
This .collect() is incompatible with passing in an existing t2g, and superfluous given the .collect() after the kb ref call.
Fixed some issues in In the test run, kb exits with a SIGKILL error: https://github.com/nf-core/scrnaseq/actions/runs/7876828275/job/21493304177#step:5:93. This is actually likely related to the binary issue in pachterlab/kallisto#411. |
t1c and t2c are required to run nac and lamanno count workflows. Since they are already defined when running kb ref, they cannot be easily defined through ext.args when using an existing reference
scrnaseq does not actually use kallisto_gene_map anywhere. Changed kallisto icon from a salmon to a rainbow (there is no bear icon, but bustools uses the Pink Floyd-esque prism image, so it seems most appropriate)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hi @gennadyFauna,
thanks for working on this! I'd like to get in #291 before we merge this to ensure everything is properly tested.
You can request to be added to the nf-core github organization in the #github-invitations
channel - then github actions should run for you without anyone approving every single time.
'https://depot.galaxyproject.org/singularity/kb-python:0.27.2--pyhdfd78af_0' : | ||
'biocontainers/kb-python:0.27.2--pyhdfd78af_0' }" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Once everything works, please update the modules in nf-core/modules
and re-install them into the pipeline using nf-core modules update
. We don't allow modules in the pipelines to diverge from the central version stored in modules
.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thank you, I have put in a request.
I appreciate the clarification regarding the nf-core/modules
source – so I would need to make changes in https://github.com/nf-core/modules/tree/master/modules/nf-core/kallistobustools, then pull them in.
If I understand correctly, these module essentially have the input/output conventions and a link to the docker image, which does not actually work.
Is there any way to rebuild the binary for the kallisto dependency?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Exactly. Whenever possible, we try to use the docker images provided by Biocontainers, which are derived from the bioconda recipe. So if the container is broken, it would be ideal to make it work there. If this is not possible for whatever reason, we can also use custom docker images in nf-core/modules
.
To just rebuild, you can make a PR to bioconda recipes that increases the build number: https://github.com/bioconda/bioconda-recipes/blob/2094e1c611eedfe500a145bd442640f0440f68a5/recipes/kb-python/meta.yaml#L14
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The bioconda recipe has been fixed. I have opened a PR at nf-core/modules#4981; this should be complete relatively soon. In the meantime, I have updated the tests and adjusted the downstream mtx operations.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@grst the nf-core module PR is merged and I have updated the kb modules in the PR branch. It still fails linting due to other files – not sure what to do; would love any suggestions. Running nf-core lint
locally does not give any failures.
Maybe it is possible to somehow inject them with ext.args, but this is a fairly standard pipeline option, and should probably be exposed.
this is not a real snapshot
mtx should take in nascent/ambiguous/mature matrices. The genes.names txt contains gene names, which are typically what is desired (e.g. cellranger has names rather than IDs)
@nf-core-bot fix linting |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM, thanks for working on this!
Don't worry about the linting failures. This is related to an update in the nf-core template that has not been yet integrated into the scrnaseq pipeline. |
@nf-core-bot fix linting |
Starting work on bumping up kb version to 0.28.2. The
nucleus
andlamanno
workflows are obsolete as of 0.28.0, so the interface for generating count matrices needs to be updated.PR checklist
nf-core lint
).nextflow run . -profile test,docker --outdir <OUTDIR>
).nextflow run . -profile debug,test,docker --outdir <OUTDIR>
).docs/usage.md
is updated.docs/output.md
is updated.CHANGELOG.md
is updated.README.md
is updated (including new tool citations and authors/contributors).