Hi all, Perhaps a straightforward question, but I couldn't find an obvious answer. Is the pipeline supporting LRT as well as Wald for DESeq2? I see it's available as an option in

--deseq2_test

, but wasn't sure if this was functional? 🙂

What happens if you try it?

Hi Jonathan, I gave it a go, for reference, my contrasts looks something like this,

id,variable,reference,target,blocking
0h_vs_1h,timepoint,0h,1h,donor_id
0h_vs_3h,timepoint,0h,3h,donor_id
0h_vs_6h,timepoint,0h,6h,donor_id
0h_vs_9h,timepoint,0h,9h,donor_id
0h_vs_12h,timepoint,0h,12h,donor_id
0h_vs_15h,timepoint,0h,15h,donor_id
0h_vs_18h,timepoint,0h,18h,donor_id
0h_vs_21h,timepoint,0h,21h,donor_id
0h_vs_24h,timepoint,0h,24h,donor_id

I essentially have nine timepoints per patient, and I'm comparing each to 0 hours as a reference and using the donor as a blocking variable, I get an error on two processes:

NFCORE_DIFFERENTIALABUNDANCE:DIFFERENTIALABUNDANCE:DESEQ2_DIFFERENTIAL
NFCORE_DIFFERENTIALABUNDANCE:DIFFERENTIALABUNDANCE:DESEQ2_NORM

In both the error based on the

.command.log

is:

Error in DESeq(dds, test = opt$test, fitType = opt$fit_type, minReplicatesForReplace = opt$min_replicates_for_replace,  : 
  likelihood ratio test requires a 'reduced' design, see ?DESeq
Execution halted

I reckon I need to provide a reduced design in the contrasts file? although I'm not sure exactly what that would look like as as far as I can tell it's setup more for pairwise testing? Thanks again for the response and any help in advanced, let me know if you need anything additional 🙂

There's a lot work going on on the design front right now, folks other than myself have made a lot of effort and can probably explain better. But the pipeline is in flux right now, and you probably need to wait for a release.

@Jonathan Manning thanks very much, assumed that was the case but good to have it confirmed! 🙂