Jim, thanks a lot for the information. I'll try it. --George -----Original Message----- From: James W. MacDonald [mailto:jmacdon@med.umich.edu] Sent: Wednesday, April 18, 2007 4:10 PM To: Tseng, George C. Cc: bioconductor at stat.math.ethz.ch Subject: Re: [BioC] Annotation of U95av2 array Hi George, I would use Entrez Gene IDs to do the matching. You could also use the mappings that Affy provide. http://www.affymetrix.com/support/technical/byproduct.affx?product=hg- u133-plus I have never used them, but they may well be useful. Best, Jim Tseng, George C. wrote: > Hi Jim, > > Your clarification is very helpful. Then when we try to match genes > from two types of arrays (say U95 and U133), what would you > recommend? We originally thought Unigene ID would be a good choice > but it would become difficult if one probe set maps to multiple IDs. > Can you advise? Sorry, it may be a dumb question but I'm from > statistics background. > > Thanks. > > George > > -----Original Message----- From: James W. MacDonald > [mailto:jmacdon at med.umich.edu] Sent: Wednesday, April 18, 2007 9:51 > AM To: Tseng, George C. Cc: bioconductor at stat.math.ethz.ch Subject: > Re: [BioC] Annotation of U95av2 array > > Hi George, > > Please don't take list conversations off-list. The list archives are > intended to be a source of information, and on the off chance that I > might say something useful, it would be nice if people could find > this later. > > As to your question, as I said below, we just map things from Entrez > Gene to the other annotation sources, so whatever Entrez Gene says, > we report. So if I grep out some probeset ID that maps to multiple > UniGene IDs, I might get something like 35566_f_at, which maps to 5 > UG IDs. > > Now if I get the Entrez ID (3576), go to the Entrez Gene webpage for > this ID, and scroll to the very bottom, I see five UniGene IDs that > this Entrez Gene ID corresponds to. We report four of these five, the > only difference being we report Hs.443948 instead of Hs.654584. > > This is obviously a mistake because Hs.443948 is SLC4A1 instead of > IL-8, but the hgu95av2 package was built on March 15, so maybe Entrez > Gene has corrected this mistake in the interim. > > See > http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=gene&cmd=Retrieve&d opt=full_report&list_uids=3576 > > > Best, > > Jim > > > Tseng, George C. wrote: > >> Jim, >> >> Thanks so much for your response. I have one further question. In >> your annotation in Bioconductor, a probe set can map to multiple >> unigene ID. This really confuses me. Shouldn't it be only one ID? >> >> George >> >> -----Original Message----- From: James MacDonald >> [mailto:jmacdon at med.umich.edu] Sent: Sunday, April 01, 2007 9:59 AM >> To: Tseng, George C. Cc: biocannotation at lists.fhcrc.org; Lu, >> Shu-Ya Subject: Re: Annotation of U95av2 array >> >> Hi George, >> >> Tseng, George C. wrote: >> >> >>> Dear Dr. MacDonald and other Biocore Data Team members, >>> >>> I'm using your array annotations from Bioconductor in my research >>> and I teach it in my microarray course as well. It is indeed a >>> great tool for our data analysis and methodological development. >>> Recently we're working on a meta-analysis research project to >>> incorporate information from multiple data sets. My student took >>> the Unigene ID annotations in all the U95av2 probes and compared >>> with the result obtained from the Affymetrix website (the batch >>> search in NetAffy). Among the 9704 probes annotated in >>> Bioconductor, 724 probes were annotated completely differently in >>> NetAffy. >>> >>> My question is: Do you obtain your Unigene ID annotation from >>> Affymetrix database or other source? NetAffy annotations always >>> have one Unigene ID to a probeset while your annotationis can >>> have many. Can you give us some detail about your annotation >>> procedure? >> >> >> Nianhua Li makes the annotation packages, so she would be the final >> trusted source. >> >> In the past, the process was to map Affy ID to Entrez Gene ID using >> the annotation files that Affy supply on their website. We then >> use AnnBuilder to do the mappings from Entrez Gene to all other >> annotation sources, so it is not inconceivable that we would have >> different UniGene IDs for a given probeset. >> >> In my experience, the BioC annotations are more up to date and >> accurate than what Affy supply either on Netaffx or in their >> annotation files. This is based on blatting the probe sequences. >> >> Best, >> >> Jim >> >> >> >> >>> Thanks! >>> >>> George >>> >>> ============================================ George C. Tseng >>> Assistant Professor Dept of Biostatistics and Human Genetics, >>> University of Pittsburgh http://www.pitt.edu/~ctseng, >>> 412-624-5318 ============================================ >> >> > > -- James W. MacDonald, M.S. Biostatistician Affymetrix and cDNA Microarray Core University of Michigan Cancer Center 1500 E. Medical Center Drive 7410 CCGC Ann Arbor MI 48109 734-647-5623 ********************************************************** Electronic Mail is not secure, may not be read every day, and should not be used for urgent or sensitive issues.