4.7 years ago by
On Fri, Nov 22, 2013 at 4:53 PM, Rohan [guest]
> I would like to get the data for all the genes in the form of Gene
> Symbols/Gene ids's mapped to GPL/GSE/GSM/GDS metadata.
> I have used GEOmetadb package to get this metadata,however I am not
> to find a way to extract all this metadata mapped to genes.
> Is their any way GEOquery bioconductor package be used for this?
Good question. It has a long-winded answer.
The GEO platform (GPL) is the only GEO entity that stores any
about gene identity. Other entities (GSM, GSE, GDS) are linked to GPL
by an ID column. So, to get information about the genes represented
experiment, we need to look at GPL records. GPL records come in two
flavors, the submitter-supplied flavor and the so-called "Annotation"
that has been curated by NCBI GEO. You'll need to focus on the
GPL since those are the ones with a standard "Gene ID" column in all
them. The "Annotation" GPLs are only generated for data sets that
been curated by NCBI GEO, namely the GDS records. So, we need to get
distinct GPL records associated with GDS and these will be the entire
of "Annotation" GPLs. Using GEOmetadb (assuming you have already made
annotgpl = dbGetquery(con,"select distinct GPL from gds")
Now, annotgpl contains the accession numbers (GPL IDs) for all the
Annotation GPLs. You can use these GPL IDs to relate each GPL to GSM,
and GSE records.
How do you get the information about what genes are on each GPL,
You'll need to use GEOquery for that.
gpl = getGEO(annotgpl[1,1],AnnotGPL=TRUE)
gpl is now a GPL object and we can use the Table method to get a data
and grab the Gene ID (which is an Entrez Gene ID):
geneids = Table(gpl)[,'Gene ID']
Now, you have the Entrez Gene IDs for all features on the platform and
can associate those with all the GSM, GDS, and GSE records attached to
GPL. If you loop over all the GPLs in the annotgpl data frame, you'll
the information you want, I think.
Unfortunately, this is not a complete answer because it does not
the submitter-supplied GPLs that do not have any Annotation GPL
(since NCBI GEO do not curate everything). The submitter-supplied
not have a standard vocabulary for what is include in the columns of
GPL, so there is not an easy way to automate processing as above.
Hope that helps.
[[alternative HTML version deleted]]