Trouble Linking Gene_Ids From *Gtf File To Assembled Transcripts In Cufflinks

Question

[User 4409]

I am having trouble getting Cufflinks to report my transcripts using the accession numbers provided by my *.gtf file. The 9th column of my *gtf file looks like this: geneid "ACYPI006883"; transcriptid "ACYPI006883"; gene_name "ACYPI006883"; so I believe that the *gtf file is properly formatted.

I have run with both the -g and -G options. When I run the -G option, it lists the transcripts with the gene ids and all the FPKM = 0. Nothing else is reported. When I run the -g option, it lists the named genes as FPKM = 0 (as with the -G option) and all the CUFF_ID genes have FPKM numbers. It's almost as if Cufflinks is telling me that it cannot match any of the assembled transcripts to my *gtf file.

I have also tried running Cuffcompare using the -r option, thinking that it might be able to link up the accession numbers, but it doesn't.

Any suggestions?

Answer 1

[Aswarren]

This has happened to me when the replicon/ contig IDs from the GTF do not match the FASTA header IDs from the indexed genome. Have you checked that?

[User 4409]

My fasta genome scaffold names are not the same as the first column of my gtf file. Should these be exactly* the same?

[Aswarren]

In my experience yes. I believe an exact string match is how cufflinks figures out which annotations belong to which scaffold/replicon.