Hi Jim, sorry for the very late reply, and thanks a lot for your detailed answer. This makes perfectly sense to me! best regards, arne "James W. MacDonald" <jmacdon@med.umich.edu> 12/21/2010 04:30 PM To arne.mueller@novartis.com cc bioconductor@stat.math.ethz.ch Subject Re: [BioC] affymetrix annotation Hi Arne, Aha! I didn't know there were two IDs associated with this probeset (although I should have thought of that, or you know, looked it up...my bad). So, here's the deal. When there are multiple IDs associated with a given probeset, it is difficult to say which one is correct. Or put another way, it is difficult to know exactly what the probeset is actually measuring. In the past, what we did was take the first Entrez Gene ID as the correct one, and ignore the rest. This of course is suboptimal, as we are ignoring information. Recently, we decided to mask these questionable mappings, and only expose the mappings that have unique probeset --> Entrez Gene mappings, but also give a way for people to get all the questionable mappings as well. If you want to see all the one-to-many mappings, you can use the infamous toggleProbes() function thusly: > suppressMessages(library(mouse4302.db)) > a <- "1419590_at" > z <- toggleProbes(mouse4302ENTREZID, "all") > get(a,z) [1] "100047700" "13094" > q <- toggleProbes(mouse4302SYMBOL, "all") > get(a,q) [1] "LOC100047700" "Cyp2b9" Best, Jim On 12/21/2010 6:15 AM, arne.mueller@novartis.com wrote: > Hello, > > thanks Jim for your reply. The org.Mm.eg.db::org.Mm.egSYMBOL has the gene > symbol for the probeset of interest, it's just mouse4302 that's missing > it. > > I've just checked in an older netffx annotation files (older than 2007, > 2009, and the current one from 2010). The probeset was annotated already > in 2007, but it was always assigned to two Entrez Gene Ids: 100047700 /// > 13094, the 2nd one seems to be the correct Id and the first one was > removed from entrez gene Oct 28th 2010 (flagged "withdrawn"). Note that > the first Id is the withdrawn one ... If the bioc annotation pipeline > comes across ambiguous annotation, that it cannot resolve, does it ignore > the annotation (since there is no unique gene Id available), and therefore > sets the entrez gene id to NA? > > Note, the gene id in entrez gene was flagged "withdrawn" in October this > year, and the date of the Entrez used in mouse4032 is September 2010 (i.e. > the entry was still valid in entrez gene - and therefore ambiguous ...). > > I'm asking because I found I found this gene interesting form some reason > and it was annotated as Cyp2b9 in our gene expressino software (using > netaffx) but not annotated at all in one of my post-processing pipelines > that uses the mouse4302.db from bioc. > > kind regards, > > Arne > > ps: the probeset I'm talking about is 1419590_at > > > > > "James W. MacDonald"<jmacdon@med.umich.edu> > 12/20/2010 09:58 PM > > To > arne.mueller@novartis.com > cc > bioconductor@stat.math.ethz.ch > Subject > Re: [BioC] affymetrix annotation > > > > > > > Hi Arne, > > On 12/20/2010 11:26 AM, arne.mueller@novartis.com wrote: >> Hello, >> >> I was wondering where the mapping from affy probeset Ids to genes >> (EntrezGene) is coming from (package mouse4302.db). >> mouse4302.db::mouse4302_dbInfo lists all data sources (URLs) However, >> what's the "original" link between an affy probeset and an EnterzGene > Id, >> is it Netaffx? My problem is that I've an "interesting" probeset that is >> annotated with an EntrezGene in Netaffx but the mouse4302SYMBOL map >> contains NA for this probeset Id. >> >> I'd be happy if someone could point me to some documentation how this >> annotation/mapping is derived or let me know what I'm doing wrong ...? > > I doubt you are doing anything wrong. This may have changed a bit from > the last time I talked to Marc, but in the past we needed a 'primary' > mapping to start with, and chose the Affymetrix probeset ID --> Entrez > Gene mapping that you can get from the latest Affy annotation file. > > We assumed this to be the truth, and then did all the other mappings > from Entrez Gene. The problem with this is that we are by definition > using more dated information than you will typically get from Netaffx. > So it is quite likely, given the fluid nature of this field, that some > probesets will have different data in Netaffx than what you can get from > the annotation packages. > > As an aside, does this probeset have an Entrez Gene ID in the annotation > package? > > Best, > > Jim > > > > >> >> thanks a lot for your help, >> >> arne >> [[alternative HTML version deleted]] >> >> _______________________________________________ >> Bioconductor mailing list >> Bioconductor@r-project.org >> https://stat.ethz.ch/mailman/listinfo/bioconductor >> Search the archives: > http://news.gmane.org/gmane.science.biology.informatics.conductor > -- James W. MacDonald, M.S. Biostatistician Douglas Lab University of Michigan Department of Human Genetics 5912 Buhl 1241 E. Catherine St. Ann Arbor MI 48109-5618 734-615-7826 ********************************************************** Electronic Mail is not secure, may not be read every day, and should not be used for urgent or sensitive issues [[alternative HTML version deleted]]