[BioC] pathview puzzle
Luo Weijun
luo_weijun at yahoo.com
Wed Aug 28 20:44:11 CEST 2013
Hi Oleg,
I just update pathview package so it can process and analyze data labeled with KEGG gene IDs other than Entrez Gene. It turns out that this issue affects many other species too. So with this update, you can literaully work with all ~2300 (and more forth-coming) KEGG species data with pathview now. I’ve also added new content with working examples on KEGG species and Gene ID usage in page 14-16 of the vignette. Notice that you need to specified gene.idtype="KEGG" when calling pathview.
I’ve posted the new package to R-forge. You should be able to access it in the next few hours at http://r-forge.r-project.org/R/?group_id=1619. Just install it follow the instruction there. The Bioc version will also be updated in the next 1-2 days: http://bioconductor.org/packages/devel/bioc/html/pathview.html.
Let me know how that works or if you have questions. HTH.
Weijun
--------------------------------------------
Subject: Re: [BioC] pathview puzzle
To: Bioconductor at r-project.org, "Oleg Moskvin" <moskvin at wisc.edu>
Date: Friday, August 23, 2013, 9:53 PM
Hi Oleg,
Thanks for the note. This is indeed a problem I didn’t
realize previously! KEGG uses Entrez Gene ID for all other
model organisms I’ve checked.
I am working on a generic fix (not only for E coli but other
species with similar situation) and will incorporate that
into the development version of pathview soon. Will keep you
posted.
Thanks for pointing this out.
Weijun
--------------------------------------------
On Fri, 8/23/13, Oleg Moskvin <moskvin at wisc.edu>
wrote:
Subject: Re: [BioC] pathview puzzle
To: Bioconductor at r-project.org,
Date: Friday, August 23, 2013, 12:19 PM
Hi Weijun,
Thank you for the response.
The problem seems to be deeper than that and is connected
to
special handling of a particular species - E.coli - by
KEGG.
I looked into the pathview() code and here is what I see:
1) gene.data is remapped internally via mol.sum() to have
ENTREZ IDs;
2) remapped gene.data is used by node.map() to map onto
KEGG
nodes using node.data
3) the node.data used in (2) was originally extracted from
the KEGG XML by node.info()
The above route implies that the "name" entries in the
KEGG
XML of type="gene" have "speciesID:ENTREZ" format...
And in the case of E.coli this doesn't hold true! See the
examples of XML entries for H.sapience and E.coli from my
yesterday's message (below).
In fact, in KEGG XML for E.coli "gene" records b-numbers
are
used as IDs!
So, for the cases like that, when KEGG fails to be
consistent in the supplied XML structure, one may suggest
introducing an "id.bypass" option to pathview() which will
take the gene.data as is (with the IDs supplied by user
that
match KEGG XML ids; for example, b-numbers), and pass this
directly to the step 3 (node matching).
Thanks!
Oleg
On 08/22/13, Luo Weijun wrote:
> Hi Oleg,
> You are right, the problem is due to ID type
inconsistency.
> You have to specify gene.idtype when calling pathview
function, if your gene id type is not Entrez Gene. I
don’t
think b-numbers are recognized for sure. For your gene
name
example, if you mean official gene symbols by “gene
name”, you should specify gene.idtype="SYMBOL" (lower
case
is fine):
> eco2.out <- pathview(gene.data =
T2.CEBF095.crt115.ASCH.DROP3.rel.gn, pathway.id = "02010",
gene.idtype="SYMBOL", out.suffix = "T2ACSH", species =
"eco", kegg.native=TRUE)
On 08/22/13, Oleg Moskvin wrote:
>
> <entry id="2" name="hsa:51343" type="gene"
> link="http://www.kegg.jp/dbget-bin/www_bget?hsa:51343">
> <graphics name="FZR1, CDC20C, CDH1, FZR, FZR2,
HCDH,
HCDH1" fgcolor="#000000" bgcolor="#BFFFBF"
> type="rectangle" x="919" y="536" width="46"
height="17"/>
> </entry>
>
>
> <entry id="4" name="eco:b1513" type="gene"
> link="http://www.kegg.jp/dbget-bin/www_bget?eco:b1513">
> <graphics name="lsrA" fgcolor="#000000"
bgcolor="#BFFFBF"
> type="rectangle" x="339" y="1882" width="46"
height="17"/>
> </entry>
More information about the Bioconductor
mailing list