introduction to GenMapp and some newbie questions

[Stanley]

Hi, I am starting to use genmapp.

Couple of questions:
a) How do I convert kegg pathways from http://www.genome.jp/kegg/kegg4.html
to mapp format for import into GenMAPP ?
b) Among the gene identifiers, which ones are most commonly used and
updated ? I am dealing with multiple platforms, and need to convert
their gene symbols to those recommended.
c) my expression data has about 20000 rows and 100 columns, any
experiences on how long the processing may take on a typical 2GHz dual
core CPU with 2GB ram ?

[GenMAPP Support]

Hi Stanley,

I have answered your questions below.

> a) How do I convert kegg pathways fromhttp://www.genome.jp/kegg/kegg4.html

> to mapp format for import into GenMAPP ?

There is no way of automatically converting KEGG pathway content
currently. We are working on a converter tool that would allow us to
periodically convert all KEGG pathways for inclusion into our
distributed archives. This should be ready later this summer, at which
point we will update our archives. For now, I'm afraid the only way to
convert KEGG content would be to hand-draw the pathway. If you plan to
do this, I would recommend drawing your pathway at WikiPathways
(www.wikipathways.org), which is where we get our pathways.

> b) Among the gene identifiers, which ones are most commonly used and
> updated ? I am dealing with multiple platforms, and need to convert
> their gene symbols to those recommended.

I would recommend annotating your data with either Entrez gene IDs or
Ensembl. These seem to be the most stable identifiers. Depending on
the species your data is from, you could also use species-specific
IDs, such as MGI, RGD, SGD etc.

> c) my expression data has about 20000 rows and 100 columns, any
> experiences on how long the processing may take on a typical 2GHz dual
> core CPU with 2GB ram ?

In terms of import to GenMAPP, I would guess that a dataset of that
size would take about 10 minutes. For downstream MAPPFinder analysis
analysis, the times go up considerably. If you are interested in GO
over-representation analysis you might consider using GO-Elite (http://
www.genmapp.org/go_elite/go_elite.html), which is much faster than
MAPPFinder, and also includes additional analysis options.

Regards,

Kristina

[Breedbeeldtelevisie]

I noticed that the Gene Ontology Mapp files originate from 2005. Are
there any plan to update those? And is there a tool to convert Gene
Ontology to MAPP?

Regards,

Maarten

[Nathan Salomonis]

There are no current plans to update these MAPPs in GenMAPP version 2,
however, MAPPFinder uses a more recent Gene Ontology database for over-
representation analysis in GenMAPP. The 2005 GO MAPPs are a sub-set
of GO terms with a minimum number of genes specifically intended to be
used for MAPPSet Export. An alternative to this is to run MAPPFinder
and export all MAPPs meeting a given criterion in MAPPFinder and then
perform a MAPPSet export on these MAPPs. Have you tried MAPPFinder
yet (GO associations from 2007)?

Best,
Nathan Salomonis (GenMAPP Dev Group)