Reference annotation file (gtf) for Nile Tilapia (Oreochromis_niloticus)

[[email protected]]

I want to do differential gene expression analysis on some Nile Tilapia RNA-Seq data using the Cufflinks-Cuffdiff method. I aligned the sequence reads to the UCSC Nile Tilapia genome with >80% mapping efficiency. I downloaded the annotation GTF file from Ensembl and converted it using the "Make ensembl GTP compatible with Cufflinks" work flow. This workflow adds "chr" in front of the chromosome number. Then I ran cufflinks on the paired-end mapped data with either the original and converted GTF file. However, the counts were 0 for all gene and transcript expression. I noticed the the chromosome names of the Tilapia gtf file are strange. They are "GL831133.1" instead of 1,2,3. I also got zero counts using HTseq-count with either the original and converted GTF file. Does anyone know a good reference annotation file that work for Tilapia? Thanks!

[dpryan]

Ensembl hasn't bothered to include oreNil2, which is what you got from UCSC. If UCSC doesn't provide an annotation then perhaps none exists.

[GenoMax]

Looks like Nile tilapia is available in UCSC Table Browser. You should be able to get a GTF file (which should match your sequence) from there.

Clade (Vertebrate) --> Genome (Nile Tilapia) --> group (Genes and Gene Predictions) --> Track (Choose one you want, RefSeq is available) --> Region (Genome) --> Output format (GTF) --> Provide a file name (e.g. tilapia_annot.gtf) --> Save/Use with HTseq-count of featureCounts.

[[email protected]]

Yes, it turns out that the UCSC genome is OreNil2.0 and the ensembl annotation is OreNil1.0. I downloaded the ensembl tilapia genome (OreNil1.0) and did the tophat alignment again. This time, I could run cufflinks and HTseq with the ensembl annotation.