Hisat2 Cufflinks Cuffmerge cuffdiff problems

Question: Hisat2 Cufflinks Cuffmerge cuffdiff problems

2.2 years ago by

Hello

I am working with cufflinks programs but I have different problems.

1.- Fist I used only cuffdiff without cufflinks, since I have D.melanogaster transcripts and in theory I can do that because is a well annotated organism. But in the results I have the cds.exp.diff file empty Why?

2.- After that I use Cufflinks (using the genome.fa and genome.gff options) and I have the transcripts.gtf outputs but also have message like that "Error (GFaSeqGet): end coordinate (1050) cannot be larger than sequence length 1049". It is importan? or I can omit that?

3.- I run cuffmerge with the "transcript.gff" and I have the merged.gff file, but also have again messages like Error (GFaSeqGet): end coordinate (2921) cannot be larger than sequence length 2916.

4._ I run cuffdiff using the merged file, the program run and I have my files but some of them are empty as CDS files

What is happening. I'm going crazy

HELP please

cuffdif cuffmerge cufflinks rna-seq hisat2 • 1.2k views

ADD COMMENT • link •

modified 2.2 years ago by Jennifer Hillman Jackson ♦ 25k • written 2.2 years ago by vm.higareda • 0

2.2 years ago by

Jennifer Hillman Jackson ♦ 25k

United States

Jennifer Hillman Jackson ♦ 25k wrote:

Hello,

Few things to troubleshoot:

The reference annotation probably contains overhanging coordinates. This could be due to a reference genome mismatch problem: https://wiki.galaxyproject.org/Support#Reference_genomes
Empty data from this tool indicates that the reference annotation is missing key attributes (tss_id, p_id, and optionally gene_name). I suggest trying again using the mm10 iGenomes version as a test: http://support.illumina.com/sequencing/sequencing_software/igenome.html. To use, download the tar file locally, unpack, and upload just the genes.gtf file to Galaxy.
If you are using the reference annotation directly from Flybase, it may have a genome mismatch issue when using it with the UCSC mm10 genome or have other content that is incompatible (or have incomplete attributes) when used with this tool package, resulting in certain output fields being reported as empty. You could use the exact genome they use as a Custom reference genome (for all steps, including mapping). They have made specific changes to enable the annotation work better with Cuff* tools, but checking in with them and reporting issues encountered will help them make it even better. They are the definitive source to learn about known issues they are working on, along with other advice for using their annotation with RNA-seq applications (this one and others).

For reference, the manual for these tools has much information about expected/utilized input content, including input reference annotation attributes. This is the link for Cuffdiff, but also see others: http://cole-trapnell-lab.github.io/cufflinks/cuffdiff/

Best, Jen, Galaxy team

ADD COMMENT • link modified 2.2 years ago • written 2.2 years ago by Jennifer Hillman Jackson ♦ 25k

Similar posts • Search »