[galaxy-user] Cuffdif output NOTEST

[Meike.Latz]

Dear all,

I have Illumina RNAseq data and want to look for differences in gene expression between male and female rats and transgenic vs. wildtype; for each condition I have triplicates. I mapped with TopHat for Illumina, using the reference genome rn5 and default settings. I did Cuffdiff afterwards and used the GTF-file Rattus norvegicus.Rnor_5.0.72.gtf as transcript. As result I got no significant changes in expression and it always says "NO TEST" (or sometimes "LOW DATA"). I found that lowering or raising the -c option might control the cuffdiff behavior, but I do not know what it is and how to control it. Can you explain me how to do it? And would you advise me to use another transcript for Cuffdiff? And when yes, where can I get it?

Lots of questions. Thanks in advance for your answer.

Best,
Meike
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.bx.psu.edu/pipermail/galaxy-user/attachments/20140408/d1fd255e/attachment.html>

[Jennifer Jackson]

Hi Meike,

Sorry for the delay. The Cuffdiff manual (http://cufflinks.cbcb.umd.edu)
has help for interpreting "NO TEST/LOW DATA" results, and for adjusting
the "-c option" - these are most often related to low coverage and/or
fragmented transcripts.

I am wondering, did you run Cufflinks to assemble the data first? This
wiki section has help for recommended protocols (and includes links back
to the tool's site above):
https://wiki.galaxyproject.org/Support#Interpreting_scientific_results
See -> Tools on the Main server: RNA-seq

As explained, it could be that there is a mismatch between the reference
annotation file and your mapped data. The gtf file contains chromosome
identifiers using Ensembl's nomenclature, while the built-in reference
genome used for mapping "rn5" is sourced from UCSC and uses their
nomenclature. These are formatted differently. An exact match is
required between identifiers or the annotation will be effectively be
ignored. Often adding a "chr" to the start of the Ensembl chromosome
name will resolve the match, but not always. The source for "rn5" was
here: http://hgdownload.cse.ucsc.edu/goldenPath/rn5/bigZips/

iGenomes is the preferred reference annotation source, due to the
inclusion of all the attributes specifically used by the Cuffdiff tool
and how these are created with specific data sources in mind (adjusted
for their identifier nomenclature). This build is not yet available, but
rn4 is: http://cufflinks.cbcb.umd.edu/igenomes.html

Checking protocol and that the identifiers are a match are the first
steps. Examine parameter tuning after.

I didn't find this data in any of the histories you already shared, but
the above help will resolve/explain most issues or results from this
pipeline.

Best,

Jen
Galaxy team

Post by Meike.Latz
Dear all,
I have Illumina RNAseq data and want to look for differences in gene
expression between male and female rats and transgenic vs. wildtype;
for each condition I have triplicates. I mapped with TopHat for
Illumina, using the reference genome rn5 and default settings. I did
Cuffdiff afterwards and used the GTF-file Rattus
norvegicus.Rnor_5.0.72.gtf as transcript. As result I got no
significant changes in expression and it always says "NO TEST" (or
sometimes "LOW DATA"). I found that lowering or raising the -c option
might control the cuffdiff behavior, but I do not know what it is and
how to control it. Can you explain me how to do it? And would you
advise me to use another transcript for Cuffdiff? And when yes, where
can I get it?
Lots of questions. Thanks in advance for your answer.
Best,
Meike
___________________________________________________________
The Galaxy User list should be used for the discussion of
Galaxy analysis and other features on the public server
at usegalaxy.org. Please keep all replies on the list by
using "reply all" in your mail client. For discussion of
local Galaxy instances and the Galaxy source code, please
http://lists.bx.psu.edu/listinfo/galaxy-dev
To manage your subscriptions to this and other Galaxy lists,
http://lists.bx.psu.edu/
http://galaxyproject.org/search/mailinglists/