For what it's worth, there are different ways to estimate the log-additive probe level model of RMA. The median polish estimator is known to have some flaws, e.g. it can oscillate between two estimates. The robust linear model M-estimator has nice properties, also gives estimates of standard errors. Since several years, Ben Bolstad provides a very solid and fast implementation in the preprocessCore package, which also handles probe-by-probe missing values and probe-by-probe weights. See for instance help("rcModelPLM", package="preprocessCore") I highly recommend to use that for estimating this PLM model instead of median polish. /Henrik On Thu, Dec 2, 2010 at 5:02 PM, Mark Cowley <m.cowley at="" garvan.org.au=""> wrote: > Hi Christian, > That's VERY interesting. > I've observed identical expression levels across probesets in newer ST > arrays, (& now that the authors mention it, probably more frequently > with odd numbers of arrays), but I can't say i've ever noticed the > phenomena in the older 3' arrays. > > Does your implementation of tRMA also apply to oligo/ST arrays? > > cheers, > Mark > > On 03/12/2010, at 8:17 AM, cstrato wrote: > >> Dear All, >> >> Recently I found an interesting paper about problems with RMA when >> applied to small sample sets which are based on the median-polish >> algorithm, see: >> http://www.biomedcentral.com/1471-2105/11/553/abstract >> >> Since the authors offer a simple solution to the problem I thought >> it might be of interest to users that I have implemented function >> "trma()" in the development version "xps_1.11.3". >> >> Best regards >> Christian >> _._._._._._._._._._._._._._._._._._ >> C.h.r.i.s.t.i.a.n ? S.t.r.a.t.o.w.a >> V.i.e.n.n.a ? ? ? ? ? A.u.s.t.r.i.a >> e.m.a.i.l: ? ? ? ?cstrato at aon.at >> _._._._._._._._._._._._._._._._._._ >> >> _______________________________________________ >> Bioconductor mailing list >> Bioconductor at r-project.org >> https://stat.ethz.ch/mailman/listinfo/bioconductor >> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor > > > > ----------------------------------------------------- > Mark Cowley, PhD > > Peter Wills Bioinformatics Centre > Garvan Institute of Medical Research, Sydney, Australia > ----------------------------------------------------- > > > ? ? ? ?[[alternative HTML version deleted]] > > _______________________________________________ > Bioconductor mailing list > Bioconductor at r-project.org > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor >

Dear Mark, Yes, my implementation of tRMA applies to all arrays including oligo/ST arrays. BTW, I forgot to mention that the authors of the paper include the R-package "trma" as supplementary file "supp12.gz" to their paper, I am sorry for this omission. Best regards Christian On 12/3/10 2:02 AM, Mark Cowley wrote: > Hi Christian, > That's VERY interesting. > I've observed identical expression levels across probesets in newer ST > arrays, (& now that the authors mention it, probably more frequently > with odd numbers of arrays), but I can't say i've ever noticed the > phenomena in the older 3' arrays. > > Does your implementation of tRMA also apply to oligo/ST arrays? > > cheers, > Mark > > On 03/12/2010, at 8:17 AM, cstrato wrote: > >> Dear All, >> >> Recently I found an interesting paper about problems with RMA when >> applied to small sample sets which are based on the median-polish >> algorithm, see: >> http://www.biomedcentral.com/1471-2105/11/553/abstract >> >> Since the authors offer a simple solution to the problem I thought it >> might be of interest to users that I have implemented function >> "trma()" in the development version "xps_1.11.3". >> >> Best regards >> Christian >> _._._._._._._._._._._._._._._._._._ >> C.h.r.i.s.t.i.a.n S.t.r.a.t.o.w.a >> V.i.e.n.n.a A.u.s.t.r.i.a >> e.m.a.i.l: cstrato at aon.at >> _._._._._._._._._._._._._._._._._._ >> >> _______________________________________________ >> Bioconductor mailing list >> Bioconductor at r-project.org >> https://stat.ethz.ch/mailman/listinfo/bioconductor >> Search the archives: >> http://news.gmane.org/gmane.science.biology.informatics.conductor > > > > ----------------------------------------------------- > Mark Cowley, PhD > > Peter Wills Bioinformatics Centre > Garvan Institute of Medical Research, Sydney, Australia > ----------------------------------------------------- >