A puzzling warning when running Cufflinks

Question

[Ding Wenchao]

Hi, all

I'm using the Cufflinks for detecting the expression in isoform level.

The command is:

cufflinks -p 8 -o out -g Homo_sapiens.GRCh38.78.gtf -b Homo_sapiens.GRCh38.dna.primary_assembly.fa -u  Aligned.sortedByCoord.out.bam

The GTF file and genome.fa was downloaded from Ensembl ftp.

I got this warning for each line which represents a gene

Warning: could not parse ID or Parent from GFF line

I'm wondering if there are something wrong with my gtf file or anything else?

Every answer will be appreciated!

Thanks in advance!

[Mike]

The gene lines in annotation files seem to cause problems with cufflinks, you can try removing them:

Remove "gene" entries (this is what I use)
https://groups.google.com/d/msg/tuxedo-tools-users/FTKA4qozJIc/p47AwnCXxvwJ

Remove everything except "exon" entries:
https://groups.google.com/d/msg/tuxedo-tools-users/V9eI65dVzyU/loGGxs3ev1oJ

Answer 1

[Ram]

You might wanna check on the chromosome naming in the GTF and Primary Assembly files. GTF uses 1,2,... naming and I'm not sure they'd match the fasta file IDs.

[Ding Wenchao]

The chromosome names in both files are consistent. This warning occured in the initial step when "Loading reference annotation", and only for the gene line not the transcript or exon line.

Thank you all the same.

Answer 2

[andrew.j.skelton73]

Consider igenomes. http://support.illumina.com/sequencing/sequencing_software/igenome.html

They have Ensembl annotation designed for the Tuxedo packages