Uploading exon results to GEO

[MPM]

I'm getting ready to send results to GEO. I think I understand their
matrix format for expression results, but I also need to include a
matrix table for exon results.

The output from AltAnalyze, as far as I can tell, does not include a
table with transcript IDs and exon expression for each sample ID.
Rather, I see "...exon-inclusion-results.txt" which has average fold
changes relative to the control condition.

Is there a way to get the exon results for each individual sample?
What have other users submitted to GEO?

-Mike

[Nathan Salomonis]

Hi Mike,

Good question.

(option 1)
In the past, I have just submitted the individual probeset expression values
(exp. file), along with the DABG detection p-values (stats. file) that are
normally saved to the ExpressionInput directory. In addition, I include the
original CEL files to GEO. Thus, someone is able to view the expression
values of probesets reported in a paper from GEO directly or reanalyze the
CEL files in AltAnalyze, if the know the database version (e.g. EnsMart60)
(http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSM335817). As for the
alternative exon results, I have submitted the "exon-inclusion-results.txt"
file as supplemental data to the journal.

(option 2)
If you want to include the normalized intensity values for each probeset
(used to calculate splicing-index) or the FIRMA residuals you can re-analyze
your data and select the option "Export all normalized intensities" from the
"Alternative Exon Analysis Parameters" window. This will save a file to the
folder AltResults/RawSpliceData containing those values. This will include
the replicate normalized intensities (relative to your method of gene
expression).

(option 3)
One other file, which probably contains what you want is the DomainGraph
input file in AltResults/DomainGraph which does have the average
splicing-index or FIRMA fold changes and associated p-values along with the
Ensembl gene. If you want corresponding Ensembl exon identifiers, you will
just need to match this file with the file
AltDatabase/EnsMart60/Hs/Hs_Ensembl_probesets.txt which has various
annotations for each probeset.

Does this work?
Best,
Nathan

[MPM]

Nathan,

Thanks for the quick reply! I didn't think to check the
ExpressionInput directory. I think option 1 will work great for now.
I might do option 2 or 3 next time I analyze an experiment.

Mike

On Apr 14, 3:22 pm, Nathan Salomonis <nsalomo...@gladstone.ucsf.edu>
wrote: