XLOC identifiers from cufflinks/cuffmerge/cuffdiff

[AdamB]

Is there a way to stop cufflinks replacing Ensembl gene IDs from the reference gtf with its own XLOC identifiers? Or is it possible to convert all the XLOCs back to Ensembl gene IDs after running cuffmerge/cuffdiff?

[Krish_143]

Use -G option i.e GTF file for tophat and cufflinks.

[AdamB]

I should have mentioned - this is when I use the -g or -G option.

[super0925]

I'd like to go back to .gtf file of reference genome and parse the Ensembl ID. But it seems that not clever. Does anyone have better or easier idea?

[danwiththeplan]

That's not what the -G option does. It means "only quantitate against genes supplied in the GTF file". While -g means "assemble new genes as well, if there is read support, and quantitate them too".

I believe that all genes get an XLOC number in the gene_id field, but this is for cufflinks to internally be able to distinguish genes. However any genes that have a name supplied in an associated GTF/GFF file will also have a name (from the gtf file) in the gene_name field. You can see this if/when you interrogate your data in cummeRbund /R.