Hi, I've been using GCRMA and the new speedier version (1.1) gives different values than the older slower version (1.0). Looking through the bioconductor mails suggests that a few other people identified a similar problem, related to a background not being subtracted. Hopefully people are on the case, but this problem seems to have been around since April. I've been plugging GCRMA to my colleagues, who are now starting to use it, so I hope the problem can be sorted out. On a different note, what technical limitations stop Affymetrix going for much longer probes than 25 bases? The work of Naef and Magnasco, and Wu and Irizarry, highlight the limitations of Affy technology due to cross-hybridisation, when there are only 25 bases. Pushing upwards to 50 bases will reduce CH, but what other factors then come in? My understanding is that the Affy SNP chips have 25 base oligos. What is stopping these chips from also having cross-hybridisation issues? Best wishes, Harry -- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~ Dr Andrew Harrison Tel: 44 (0) 207 679 3890 Biomolecular Structure and Modelling Unit Fax: 44 (0) 207 679 7193 Biochemistry and Molecular Biology Dept. University College London Gower Street Email: harry@biochem.ucl.ac.uk London, WC1E 6BT, UK http://www.biochem.ucl.ac.uk/~harry ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~

As far as I'm aware there is no problem with gcrma at present (it returns no errors). I'm sure one of the authors may confirm this. As for the 25mers, the obvious thing to take into account is that as you increase in length it is more likely that non-homologous probes will bind as it would be more difficult to find sequences that are gene specific. HTH, Matt

What are references for this? Mike >>> "Matthew Hannah" <hannah@mpimp-golm.mpg.de> 07/21/04 12:45PM >>> As for the 25mers, the obvious thing to take into account is that as you increase in length it is more likely that non-homologous probes will bind as it would be more difficult to find sequences that are gene specific. HTH, Matt _______________________________________________ Bioconductor mailing list Bioconductor@stat.math.ethz.ch https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor

There is some changes between the older version and the newer ones. The original use of sequence information only considered GC counts in the probe sequence, while later a more sophisticated model is used that considers the position and types of base. This difference comes from our approach on using the sequence information, not from the speed. Jean(Zhijin) On Wed, 21 Jul 2004, Andrew Harrison wrote: > Hi, > > I've been using GCRMA and the new speedier version (1.1) > gives different values than the older slower version (1.0). > > Looking through the bioconductor mails suggests that > a few other people identified a similar problem, related to a > background not being subtracted. Hopefully people are on the case, > but this problem seems to have been around since April. I've been > plugging GCRMA to my colleagues, who are now starting to use it, > so I hope the problem can be sorted out. > > On a different note, what technical limitations stop > Affymetrix going for much longer probes than 25 bases? The work > of Naef and Magnasco, and Wu and Irizarry, highlight the > limitations of Affy technology due to cross-hybridisation, when > there are only 25 bases. Pushing upwards to 50 bases will reduce CH, > but what other factors then come in? > > My understanding is that the Affy SNP chips have 25 base > oligos. What is stopping these chips from also having > cross-hybridisation issues? > > Best wishes, > Harry > > -- > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~ > Dr Andrew Harrison Tel: 44 (0) 207 679 3890 > Biomolecular Structure and Modelling Unit Fax: 44 (0) 207 679 7193 > Biochemistry and Molecular Biology Dept. > University College London > Gower Street Email: harry@biochem.ucl.ac.uk > London, WC1E 6BT, UK http://www.biochem.ucl.ac.uk/~harry > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~ > > _______________________________________________ > Bioconductor mailing list > Bioconductor@stat.math.ethz.ch > https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor >

I should have said it was just a logical guess. What I meant was that if you had 2 homologous genes, obviously it is going to be harder to avoid homologous regions if you need to find 50bp versus 25bp? But this is refering to cross-hybridisation between PM and related sequences, I don't know how it would affect non- specific binding of PM to non-complementary sequences (am I right to distinguish these?). I should have said 'less-' rather than non-homologous, or just dropped the 'non-' in the initial post. Also this would only apply where there were related sequences present, but then different probe-lengths for different sequences wouldn't be ideal. Also while we're on logic another reason to consider is that with 11-20 probesets per mRNA, for short mRNAs there is already some overlap, this would be worse for longer probes, making them less independent. It would also extend the probed region further from the 3' end from where labelling occurs and so efficiency may be reduced? If you need a reference I'm sure the affy website or some of their publications would have something. Sorry for any confusion. Matt -----Original Message----- From: Michael Barnes [mailto:Michael.Barnes@cchmc.org] Sent: Mittwoch, 21. Juli 2004 19:49 To: Matthew Hannah; bioconductor@stat.math.ethz.ch Subject: Re: [BioC] GCRMA backgrounds? What are references for this? Mike >>> "Matthew Hannah" <hannah@mpimp-golm.mpg.de> 07/21/04 12:45PM >>> As for the 25mers, the obvious thing to take into account is that as you increase in length it is more likely that non-homologous probes will bind as it would be more difficult to find sequences that are gene specific. HTH, Matt _______________________________________________ Bioconductor mailing list Bioconductor@stat.math.ethz.ch https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor Hi, I've been using GCRMA and the new speedier version (1.1) gives different values than the older slower version (1.0). Looking through the bioconductor mails suggests that a few other people identified a similar problem, related to a background not being subtracted. Hopefully people are on the case, but this problem seems to have been around since April. I've been plugging GCRMA to my colleagues, who are now starting to use it, so I hope the problem can be sorted out. On a different note, what technical limitations stop Affymetrix going for much longer probes than 25 bases? The work of Naef and Magnasco, and Wu and Irizarry, highlight the limitations of Affy technology due to cross-hybridisation, when there are only 25 bases. Pushing upwards to 50 bases will reduce CH, but what other factors then come in? My understanding is that the Affy SNP chips have 25 base oligos. What is stopping these chips from also having cross-hybridisation issues? Best wishes, Harry