I still have some doubt about your experiment: you are comparing control vs treatment after 6 or 24 hour (so time series, not two control plus treatment A plus treatment B)?
As to the cuffdiff output: you should normally get the common gene name in your various files. If you don't cufflinks use its own identifiers ('tracking_id' like gene id, isoform id) for genes and transcripts (look at the format description on the cuffdiff page), which is also present in the annotation I think. You can try to merge these information (using R or databases).
It is still hideously complicated to do by hand. I have never used CummeRbund BUT looking at the manual page 6 and especially 7:
- "All of these output files are related to each other through their various tracking ids, but parsing through individual files to query for important result information requires both a good deal of patience and a strong grasp of commandline text manipulation. Enter cummeRbund, an R solution to aggregate, organize, and help visualize this multilayered dataset."
- "We now also recommend that you use both the genome and gtfFile arguments to readCufflinks(). This will allow cummeRbund to archive the transcript structure information located in the .gtf file associated with your particular cuffdiff run, as well as associate these transcripts with an appropriate genome build (e.g. 'hg19', 'mm9', etc) so as to allow for transcript-level visualizations and future integration with other external resources."
So just create a R script, after installing cummeRbund:
library(cummeRbund)
cuff<-readCufflinks(dir="/home/blubelle92/thePlaceWhereYourCuffdiffOutputIs",gtfFile="cuffmerge_out/merged.gtf",genome="TAIR10") # I know nothing on plant genetics, use the correct genome
I have the same type of issues with my own master's thesis, and juggling between a reference annotation using Ensembl gene ids, a custom made one using custom ids and genes with various official and unofficial names is just plain tedious. Cufflinks is practical and does a lot of work for you, which is sometimes a problem. Don't trust blindly the way it counts gene expression levels or isoforms, check your data using IGV.
•
written
by
cyril-cros
Did you provide the annotated GTF with cuffmerge?
Yes, I just forgot to add the altered command in the question.
The result of adding the annotated gtf file was annotated genes but with no values when run through cuffdiff.
Is that your complete cuffdiff command? If it is, I'd imagine that it's having trouble, because you're telling it to test three sample types in singlet. it should be
blah.gtf condA_rep1.bam,condA_rep2.bam condB_rep1.bam,condB_rep2.bam... if I remember rightly...