refine_VIF_output.R fails when *.vif.evidence_counts.tsv is empty

[katgorski]

Description of the bug

So there weren't any interesting events in my sample and *vif.evidence_counts.tsv was empty, only had a header. Thus I wasn't able to generate the sample report defined in the SUMMARY_REPORT process because that file was empty--in refine_VIF_output.R, the resulting col types didn't match so it wasn't able to execute the join here:

viralintegration/bin/refine_VIF_output.R

Lines 53 to 55 in 70e558a

	vif_counts <- read.table(vif_counts_filename, header = T, sep = "\t", stringsAsFactors = F, comment.char = "", quote = "")
	data <- right_join(prelim_counts, vif_counts, by = "contig")

It should be a simple fix: define the column types when reading in the table initially, but because it's not originally your script/it's from the original pipeline you adapted, I'm not sure how much you want to modify it.

Command used and terminal output

Command exit status:
  1

Command output:
  (empty)

Command error:
  Error in `right_join()`:
  ! Can't join on `x$contig` x `y$contig` because of incompatible types.
  i `x$contig` is of type <character>>.
  i `y$contig` is of type <logical>>.
  Backtrace:
      x
   1. +-dplyr::right_join(prelim_counts, vif_counts, by = "contig")
   2. \-dplyr:::right_join.data.frame(prelim_counts, vif_counts, by = "contig")
   3.   \-dplyr:::join_mutate(...)
   4.     \-dplyr:::join_rows(...)
   5.       \-base::tryCatch(...)
   6.         \-base (local) tryCatchList(expr, classes, parentenv, handlers)
   7.           \-base (local) tryCatchOne(expr, names, parentenv, handlers[[1L]])
   8.             \-value[[3L]](cond)
   9.               \-rlang::abort(bullets, call = error_call)
  Execution halted

Relevant files

No response

System information

• Nextflow version 23.10.1 build 5891
• Seqera platform to launch
• azurebatch
• nf-core/viralintegration v0.1.1 (70e558a)

[mletexier-cnrgh]

Dear katgorski
I have the same error as you. Were you able to fix the error and finalize the pipeline?

Sincerely
Mélanie

[katgorski]

Dear katgorski I have the same error as you. Were you able to fix the error and finalize the pipeline?

Sincerely Mélanie

hey @mletexier-cnrgh! so running the pipeline for my data was a one-off to examine samples from one patient. the result of "the empty tsv says there were no viruses here" was sufficient for us (we didn't expect anything) so we didn't fix the error to get to the report stage. I actually describe a potential fix in the issue and it would be pretty easy to implement, however I didn't want to fork and do it right away as it is modifying code adapted from another pipeline. getting some example data would be helpful though, the stuff i had done was deleted long ago.

a release hasn't been cut since 2023 and while there are quite a few commits in the dev branch to keep track with nf-core templates, there hasn't been a merge to main. the slack channel for this pipeline is pretty quiet too. @alyssa-ab seems to be the most active person here--if i forked and created a PR to dev, would it be looked at? (i haven't contributed to nf-core yet either, so not totally familiar with the process.)

[mletexier-cnrgh]

Thank you very much for taking the time to reply. Yes, the correction is very simple, and personally I don't want to mess with the code either ^^. I'm not very comfortable with it either.
I'll settle for these results even though I think I'll lose the final multiQC.
All the best