An embedded and charset-unspecified text was scrubbed... Name: not available Url: https://stat.ethz.ch/pipermail/bioconductor/attachments/20061214/ 119f8de4/attachment.pl

Hi Greg. It is not too surprising - there are 11-16 probes in a probeset and sometimes some of them may match other genes/transcripts than intended initially by Affy. Thus the structure of annotations is not as simple as a probeset-gene pairs. There is also a number of papers on evolving annotations - the databases like Ensembl are not stable - transcripts/ESTs come and go... That's why you can rely on annotations only in an approximate way. If you want to be sure - map (e.g. blast) the probes yourself - then you have much more control on what is targeted by the probeset or check something like our web apps : http://bioinformatics.picr.man.ac.uk/adapt/Welcome.adapt http://bioinformatics.picr.man.ac.uk/adaptnet/ Cheers, Michal -----Original Message----- From: bioconductor-bounces@stat.math.ethz.ch [mailto:bioconductor- bounces@stat.math.ethz.ch] On Behalf Of gregory voisin Sent: 14 December 2006 20:15 To: bioconductor Subject: [BioC] difference between annotation Hi Bioconductorians!!!! I work with HGU133.2.plus I would like why there is a difference between annotation got limma in and Affymetrix's annotation?????? because sometimes for a probeset given , the limma's annotation gives on information and affymetrix gives another information...... ps: (( Moreover, I add that the annotation of Pathway Assist are too different ( AROUND 15 TO 20 %)AND THE STRATAGENE'S TEAM (PRIVATE COMPANY) ARE VERY GOOD IN BUSINESS BUT NOT VERY GOOD IN RELATIONSHIP .... it will be to get a consensus , one day .....)) thanks Greg ______________________________________________________________________ _____ interface r?volutionnaire. [[alternative HTML version deleted]] -------------------------------------------------------- This email is confidential and intended solely for the use o...{{dropped}}

Hi Gregory, There are some differences between limma(bioconductor) annotations and Affymetrix annotations because bioconductor and Affymetrix use different methods to obtain the annotations. Bioconductor uses the RepresentativePublicIdentifier given by Affymetrix as the starting point for the annotations. Affymetrix currently tries to align the probes to a transcript which it calls the Annotation Transcript, and the annotations are those for that transcript. For some probe sets the different methods assign the probe set to a different gene. Also, as Michal mentioned, databases such as GenBank are constantly changing, and knowledge of the genome is constantly evolving, so you can expect to find annotation differences, sometimes of the order of 5-10%, between one release of bioconductor annotations and the next, and between one release of Affymetrix annotations and the next. Regards, Maria