This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Cufflinks output with filename prefix

Hi all,

I am using gnu parallel for cufflinks.I had question on how to give the prefix to the output files.Since I have a lot of bam samples and the output files produced are of the same name and would replace each time for the new sample.

    parallel -j $NSLOTS --xapply \
   "cufflinks -G gencode.v19.annotation.gtf {1} > {1/.}." ::: /input/*.bam

I am using this command above but it does not work.

Thanks,
Ron

next-gen rna-seq cufflinks
• 5,437 views • link
updated by Ram Ram • written by Ron Ron

Could you include an example of the input and the command you expected it to run?

• 0 views • link
written by ole.tange ole.tange

I mean the input files are of different name but cufflinks generates files transcripts.gtf, genes.fpkm_tracking, and isoforms.fpkm_tracking.So every time a sample is run,these output files with same name (transcripts.gtf, genes.fpkm_tracking, and isoforms.fpkm_tracking) are generated again. I want these output files to be named according to the sample.

e.g 508_C.bam would be named as 508_C.genes.fpkm_tracking, 508_C.transcripts.gtf.

• 0 views • link
updated by Ram Ram • written by Ron Ron

Ahh. So you want GNU Parallel to help fix a problem in cufflinks. It cannot do that directly, but you can help it by making a function that does that for your. Untested:

doit() {
  mkdir $1
  cd $1
  cufflinks -G gencode.v19.annotation.gtf "$2" > ../"$3"
  mv transcripts.gtf ../"$3"-transcripts.gtf
  mv genes.fpkm_tracking ../"$3"-genes.fpkm_tracking
  mv isoforms.fpkm_tracking ../"$3"-isoforms.fpkm_tracking
  cd ..
  rm -r $1
}
export -f doit
parallel doit {#} {} {/} ::: /input/*.bam
• 0 views • link
updated by Ram Ram • written by ole.tange ole.tange

1 answer

If you have a lot of input files that are named the same and you want the output files named uniquely, we somehow need a way make the names unique. The most obvious way would be to use {#} which is the job number:

parallel cufflinks -G gencode.v19.annotation.gtf {} '>' {/.}.{#} ::: /input/*.bam

Another idea is to append a random value (and hope for no clashes):

parallel cufflinks -G gencode.v19.annotation.gtf {} '>' {/.}.'{= $_=int(1000000*rand) =}' ::: /input/*.bam

If the full path is unique, you can simply change the / to _:

parallel cufflinks -G gencode.v19.annotation.gtf {} '>{=s:/:_:g=}' ::: /input/*.bam
• 0 views • link
updated by Ram Ram • written by ole.tange ole.tange

The file names are different but I am getting only 4 output files for all bam samples namely:

genes.fpkm_tracking
isoforms.fpkm_tracking
skipped.gtf
transcripts.gtf

I expect to have separate files for my each sample, so want to add a prefix somehow as I think I am getting the results of the last bam sample only. I tried the first command but the output is same.

• 0 views • link
updated by Ram Ram • written by Ron Ron

Can you explain this part {} '>' {/.}.{#} ::: /input/*.bam since there might be bam files which would be from different samples and each cufflink output I would like to be in different folder. How can I do that ?with the command you have mentioned?

• 0 views • link
updated by Ram Ram • written by 1769mkc 1769mkc

Log in to answer this question.

More posts like this