Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • Jan 31 2019 17:58
    jorgemachucav starred galaxyproject/tools-iuc
  • Jan 31 2019 17:45
    bebatut opened #2270
  • Jan 31 2019 16:18
    cpreviti synchronize #2267
  • Jan 31 2019 14:15
    cpreviti synchronize #2267
  • Jan 31 2019 12:42
    bernt-matthias review_requested #2269
  • Jan 31 2019 12:42
    bernt-matthias edited #2269
  • Jan 31 2019 12:41
    bernt-matthias edited #2269
  • Jan 31 2019 12:40
    bernt-matthias synchronize #2269
  • Jan 31 2019 12:13
    cpreviti commented #2267
  • Jan 31 2019 12:07
    nsoranzo commented #2267
  • Jan 31 2019 12:01
    cpreviti synchronize #2267
  • Jan 31 2019 11:21
    cpreviti synchronize #2267
  • Jan 31 2019 09:47
    cpreviti synchronize #2267
  • Jan 31 2019 09:27
    cpreviti synchronize #2267
  • Jan 30 2019 20:38
    bernt-matthias commented #2131
  • Jan 30 2019 20:19
    hepcat72 commented #2239
  • Jan 30 2019 19:50
    lparsons commented #2239
  • Jan 30 2019 18:36
    bgruening commented #2268
  • Jan 30 2019 15:23
    nsoranzo commented #2268
  • Jan 30 2019 15:23
    nsoranzo commented #2267
Lance Parsons
@lparsons
Is there a way to output a file from featureCounts that works both in DESeq2 as well as MultiQC? I'm a bit confused by the current options in version 1.6.0.2? @mvdbeek ?
Ghost
@ghost~5772e7e2c2f0db084a206e1b
Not sure, I haven't followed the discussions / updates. So the featurecounts output format has changed ?
Adapting this for multiqc while breaking deseq2 sounds like a bad deal
(if that's what happened)
Brad Langhorst
@bwlang
@nsoranzo that solves one ugliness, but i’m not sure about searching a subtree for an fai… is that likely to work reliably on other galaxys. Maybe i should just change the tool to use ths sam_indexes ref source...
Lance Parsons
@lparsons
Yeah, looking a bit deeper, it seems it was this line: https://github.com/galaxyproject/tools-iuc/pull/1664/files#diff-9c6098d25fe7aa5d7d90c3e2c9ff4aebR178
@mblue9 Have you tested this? Am I right in assuming DESeq2 doesn't work with the "MultiQC Compatible" output from FeatureCounts?
Ghost
@ghost~5772e7e2c2f0db084a206e1b
OK, you have to select tabdel_short now
but the bigger issue is these mega PRs are barely reviewable
because that change should really not have passed as everyone has to update their workflows now
Lance Parsons
@lparsons
Agreed. And worse, I can't realistically use featurecounts, deseq2, and multiqc in the same workflow?! That's not right. MultiQC and DESeq2 should take the same input. Those should be fixed, not have featurecounts effectively broken...
Hrm.
Ghost
@ghost~5772e7e2c2f0db084a206e1b
Also this is pretty suboptimal because now you've got to run featurecounts once for multiqc and once for deseq2 or drop the header before passing into deseq2.
Lance Parsons
@lparsons
Agree completely.
Ghost
@ghost~5772e7e2c2f0db084a206e1b
On top of that featurecounts doesn't start the head with #, so we can't just handle that in deseq2
Lance Parsons
@lparsons
Not sure how to fix this exactly, but it seems there should be a reasonable way to handle it. Perhaps adding the header in the MultiQC wrapper from the metadata?
That is pretty nasty trickery (and not really documented, afaik), but it seems it would keep the other tool wrappers (deseq2 and featurecounts) clean.
Ghost
@ghost~5772e7e2c2f0db084a206e1b
yeah, I think that'd be better
Lance Parsons
@lparsons
Is there an example of this anywhere you know of? Grabbing the metadata I mean (and perhaps adding it as a header line). I seem to remember seeing something like this before. Might be nice to have this as a built in option that tool authors could specify (some verion of the input file with headers as the first line..., but that's a bit too long term for this.
Ghost
@ghost~5772e7e2c2f0db084a206e1b
I suppose you can construct this from the element identifier
Lance Parsons
@lparsons
No, I don't think so, since it there could be mulitple columns, right? At least in theory.
Ghost
@ghost~5772e7e2c2f0db084a206e1b
Well that format doesn't work for deseq2 at all
Lance Parsons
@lparsons
Also, I'm curious why this works with edgr and limma-voom
That's not really great either...
Ghost
@ghost~5772e7e2c2f0db084a206e1b
It's another approach ...
Lance Parsons
@lparsons
But a bit of a bigger problem to solve...
MultiQC needs to be able to take as wide a range of outputs from FeatureCounts as possible, imho.
If DESeq2 is constrained, so be it.
Ghost
@ghost~5772e7e2c2f0db084a206e1b
So for the deseq2-style output you could write a configfile that just contains Geneid $input1.element_identifier and then < $configfile $input1 > multiqc_input.tab or sth. like that
The other DE wrappers take the header format, so that wouldn't be an issue
Lance Parsons
@lparsons
Or should we just allow DESeq2 to accept the header format? And output the header line with a leading #?
Ghost
@ghost~5772e7e2c2f0db084a206e1b
That'd be OK as well, I think
Sounds less complicated, if i think about it :D
Well, we can treat a line that starts with Geneid as a header line in deseq2, I think that'd be better than manipulating the featurecounts output
who knows who's depending on that ...
(if I ever discover a gene I'll name it Geneid though :D)
Lance Parsons
@lparsons
I'd be OK with that and then removing the DESeq2 specific output from FeatureCounts...
Ghost
@ghost~5772e7e2c2f0db084a206e1b
:+1:
Dave B.
@davebx
@lparsons goodness, I'm dense. I've been trying to integrate your enhancements of fastq_barcode_splitter into 1413, and fighting with the stock script's error message of Unknown option: idxfile
no points to anyone who can guess the very obvious error I made there
Lance Parsons
@lparsons
@davebx ;-) Let me know if I can be of help...
Dave B.
@davebx
I think I got it, see #1413 for details
Lance Parsons
@lparsons
Cool...
Lance Parsons
@lparsons
@mvdbeek @mblue9 No matter what I do, I'm unable to get a workflow working that includes featureCounts 1.6.2 as part of a subworkflow to work. I get something like: Exception: Workflow evaluation problem - failed to find output_name output_jcounts in step_outputs {'output_summary': <galaxy.model.HistoryDatasetCollectionAssociation object at 0x7efde034c1d0>, 'output_short': <galaxy.model.HistoryDatasetCollectionAssociation object at 0x7efdda5d9d90>} Any thoughts?
Also, once I create the parent workflow, I cannot go back and edit it, it just hangs trying to load the workflow....
Lance Parsons
@lparsons
And now I'm getting Exception: Could not find connections for all subworkflow inputs....
Lance Parsons
@lparsons
Well, somehow I've gotten a minimal version to at least schedule jobs (no idea what is different).
Lance Parsons
@lparsons
OK, I've found the bug. It seems that once I mark some outputs as a workflow output, the workflow becomes broken and I'm no longer able to load it. Haven't tested, but I'm guessing running it is broken as well.
Lance Parsons
@lparsons
Submitted issue galaxyproject/galaxy#6031
Lance Parsons
@lparsons
Crap. Spoke too soon. Now I can load the workflow, and even save it. BUT, any modification (includine simply auto-reflow) causes the dreaded Uncaught exception in exposed API method: error and I can't save the workflow...