On Tue, 2003-09-30 at 12:09, Ann Hess wrote: > Does expresso ALWAYS return logged data? For example, when calling the > LiWong algorithm using No expresso() does not always return log scale expression values. In fact I believe that only summary.method="medianpolish" returns log scale expression values. It should be pretty apparent by examining a few expression values to confirm that this is the case. eg do something like library(affy);library(affydata) data(Dilution) eset <- expresso(Dilution,normalize.method="invariantset",bg.correct=FALSE,pmc orrect.method="pmonly",summary.method="liwong") exprs(eset)[1:5,] which, for me returns something like 20A 20B 10A 10B 1000_at 639.2587 597.4893 665.5438 591.6256 1001_at 159.1597 149.6903 149.0489 152.0907 1002_f_at 337.3051 305.9068 312.6497 295.9938 1003_s_at 647.0458 626.9708 601.9409 592.7388 1004_at 353.2921 314.6988 318.9967 316.4813 which definitely look non-log scale to me

li wong returns original/antilog scale. originally, the idea was to return values in the scale which the model, used for the data reduction, suggests one should perform tests. so for li and wong's model (additive error) it comes back as original scale. for the medianpolish function used by rma (multiplicative error) it comes back in the log scale. because its not clear what the model is for mas5 and mas4 (avdiff) original scale is returned for consistency with their software. i believe vsn returns values in the scale they suggest one should perform test, i.e. with the variance stabilizing transformation applied. hopefully this isnt too confusing. i believe/hope the help files say what scale is used. On Tue, 30 Sep 2003, Ann Hess wrote: > > Does expresso ALWAYS return logged data? For example, when calling the > LiWong algorithm using > > expresso(data,normalize.method="invariantset",bg.correct=FALSE, > pmcorrect.method="pmonly",summary.method="liwong") > > does that return expression values on the log or original scale? > > > > On Tue, 30 Sep 2003, James MacDonald wrote: > > > Assuming you want the log of the PM data, you would do log2(pm(abatch)). > > For log2(MM), substitute mm for pm. > > > > Note also that expresso and rma return logged data, so you would not > > want to input pre-logged data to these functions. > > > > Jim > > > > > > > > James W. MacDonald > > Affymetrix and cDNA Microarray Core > > University of Michigan Cancer Center > > 1500 E. Medical Center Drive > > 7410 CCGC > > Ann Arbor MI 48109 > > 734-647-5623 > > > > >>> Stephen Nyangoma <s.nyangoma@cs.rug.nl> 09/30/03 12:50PM >>> > > Hi all, > > How can I compute the logs of the Affybatch data? > > > > Stephen. > > > > _______________________________________________ > > Bioconductor mailing list > > Bioconductor@stat.math.ethz.ch > > https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor > > > > _______________________________________________ > > Bioconductor mailing list > > Bioconductor@stat.math.ethz.ch > > https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor > > > > _______________________________________________ > Bioconductor mailing list > Bioconductor@stat.math.ethz.ch > https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor >