gff files in kmer analysis.

[shao]

Hello,

First I want to thank you for the great ribogalaxly server!
I followed the published workflow to map the ribosome
profiling data to mouse genome. Later in the kmer analysis in
ribo tools section, a gffs file is required.

Does the ribogalaxy has the gff3 related to the work flow?
If not, should I use one from ensembl or from UCSC?

Regards,

Shao

[jamesp]

Hello Shao,
we have not tried running kmer on mouse data ourselves. Anyhow the necessary gff file is not available on RiboGalaxy. So you will need to retrieve and upload the gff file yourself. An important recommendation mentioned in another post on this forum is that the genome fasta file and the gff file need to be from the same source and need to refer to the chomosome sequences in the same format . Unfortunately this requirement is not always met for different species from the same library and it is therefore difficult to give a definitive answer. The best advice we can give is to try using files from either Ensembl or NCBI and see what is the outcome.
Thanks to let us know what happens if you try this.

Regards
James

[shao]

Hi James,

Thanks for quick reply. I looked the bam files of genome mapping and chromosomes start with "chr", ensembl/NCBI gff will not work as there is no "chr" as prefix.

It is difficult to get a gff3 for ucsc annotation.

I will try to upload the ensembl and redo the genome mapping step, and then kmer.

Regards,
Shao

(06-Nov-2017 12:50 PM)jamesp Wrote:  Hello Shao,
we have not tried running kmer on mouse data ourselves. Anyhow the necessary gff file is not available on RiboGalaxy. So you will need to retrieve and upload the gff file yourself. An important recommendation mentioned in another post on this forum is that the genome fasta file and the gff file need to be from the same source and need to refer to the chomosome sequences in the same format . Unfortunately this requirement is not always met for different species from the same library and it is therefore difficult to give a definitive answer. The best advice we can give is to try using files from either Ensembl or NCBI and see what is the outcome.
Thanks to let us know what happens if you try this.

Regards
James

[shao]

Hi James,

So I reran the kmer analysis with Ensemble gff3 and a sorted bam file generated by STAR.
The gff3 and genome are for ensembl 86.

kmer gave the following error:
Error during footprinting : list index out of range

I read a thread showing the similar problem and the solution is to use the
genome from NCBI. However, ensembl is used in other part of the project and I would to
stick to the same one. Do you have an idea why it does not work for ensembl?