Dear useR, I use R 2.5.0 since it's release and so for no problem... to I had to re-process some data and I encounter a problem while trying to use the mas5 function from the affy package: > da.26may.mas5 <- mas5(da.26may, sc=300) background correction: mas PM/MM correction : mas expression values: mas background correcting...done. 12625 ids to be processed | | |Error in update.default(pbt) : need an object with call component In addition: Warning message: $ operator not defined for this S4 class, returning NULL in: object$call > traceback() 8: stop("need an object with call component") 7: update.default(pbt) 6: update(pbt) 5: .local(x, pmcorrect.method, summary.method, ...) 4: computeExprSet(afbatch, summary.method = summary.method, pmcorrect.method = pmcorrect.method, ids = summary.subset, summary.param = summary.param, pmcorrect.param = pmcorrect.param) 3: computeExprSet(afbatch, summary.method = summary.method, pmcorrect.method = pmcorrect.method, ids = summary.subset, summary.param = summary.param, pmcorrect.param = pmcorrect.param) 2: expresso(object, bgcorrect.method = "mas", pmcorrect.method = "mas", normalize = FALSE, summary.method = "mas", ...) 1: mas5(da.26may, sc = 300) > sessionInfo() R version 2.5.0 Patched (2007-04-26 r41343) i386-apple-darwin8.9.1 locale: C attached base packages: [1] "stats" "graphics" "grDevices" "datasets" "utils" "tools" "methods" "base" other attached packages: hgu95av2 hgu95av2cdf affy affyio Biobase "1.16.0" "1.16.0" "1.14.1" "1.4.0" "1.14.0" --- David Ruau Institute for Biomedical Engineering RWTH Aachen

Hi David, David Ruau <david.ruau at="" rwth-aachen.de=""> writes: > Dear useR, We're bioconductoRs here ;-) > I use R 2.5.0 since it's release and so for no problem... to I had to > re-process some data and I encounter a problem while trying to use > the mas5 function from the affy package: > > > da.26may.mas5 <- mas5(da.26may, sc=300) > background correction: mas > PM/MM correction : mas > expression values: mas > background correcting...done. > 12625 ids to be processed > | | > |Error in update.default(pbt) : need an object with call component > In addition: Warning message: > $ operator not defined for this S4 class, returning NULL in: object$call For some reason, it looks like the update method isn't dispatching properly for you. One possible workaround: what happens if you repeat the call, but specify verbose=FALSE? Does example(mas5) work for you? + seth -- Seth Falcon | Computational Biology | Fred Hutchinson Cancer Research Center http://bioconductor.org

David Ruau <david.ruau at="" rwth-aachen.de=""> writes: > verbose=FALSE or TRUE doesn't change anything > and example(mas5) give the same error... > I re-install the last R-patched from source nothing changed, I > install in despair R binary and I am still having the same problem. Hmm, how have you been instaling Biocondcutor packages? If you are using a source built R, I recommend (removing all possibly stale BioC packages and then): source("http://bioconductor.org/biocLite.R") biocLite("somePkg", type="source") After getting a good install, you can update like this: library("Biobase") update.packages(repos=biocReposList()) If you are using the CRAN binary R, then you want type="mac.binary" which is the default on OSX. I run OS X on my laptop, but use a source-built R. I have the following in ~/.Rprofile: options(repos="http://cran.fhcrc.org") options(pkgType="source") options(browser="open") if (TRUE && interactive()) { tryCatch({ source("http://bioconductor.org/biocLite.R") }, error=function(e) invisible(NULL), warning=function(w) cat("Not connected to the net\n")) } Setting the pkgType option to "source" means I can call biocLite() without remember to ask for source packages. > Can I re-install update.default() method manually? No, the problem is not exactly a wrong definition. I'm not yet convinced it is a problem in the software and suspect that you have a bad mix of versions. Let's see if we can sort that out first. + seth -- Seth Falcon | Computational Biology | Fred Hutchinson Cancer Research Center http://bioconductor.org

David Ruau <david.ruau at="" rwth-aachen.de=""> writes: > I still have the same error unfortunately. > I don't have the last gcc (gcc-4.0.1) install yet but I don't think > that's the problem... Expect problems if you use a different version of gcc to install source packages using an R built using different compilers (AFAICT, this is exactly what you are doing). > Do you know another program using this method update()? maybe I can > more information for debugging. The issue looks to be in the _dispatch_ of update not in the method itself. But since I'm not able to reproduce what you are seeing, I'm not sure what to suggest. > Here is what I did: > I cleaned the entire /Library/Frameworks/R.framework/Versions/2.5/ > Resources/library/ > then I re-installed R binary, modified my .Rprofile to install from > source. Nooo. If you install the R binary you want to install binary R packages (which is the default). In general, you should only use source packages on OS X if you built R from source... In fact, there should be no problem mixing source and binary as long as you have the same XCode Tools and gcc, etc as that which was used to build R. Furthermore, I tried mixing and matching as you have done and _still_ cannot reproduce the error. >> da.26may <- ReadAffy(celfile.path="./CEL_files/") >> da.26may.mas5 <- mas5(da.26may, sc=300) > background correction: mas > PM/MM correction : mas > expression values: mas > background correcting...done. > 12625 ids to be processed > | | > |Error in update.default(pbt) : need an object with call component > In addition: Warning message: > $ operator not defined for this S4 class, returning NULL in: > object$call It would be better to show us the failed example, since that is something we can all try to reproduce easily. We've found at least one bug so far in affy since verbose=FALSE is broken. I get: > eset <- mas5(affybatch.example, verbose=FALSE) 150 ids to be processed | | |####################| So you really meant it when you said it didn't change anything! I can run the mas5 example without error in the UTF-8/en_US and C locales. I also tried rerunning after installed the latest affy package from source (1.14.1 is not currently available as a Mac binary). But everything is working. :-\ In a new R session, I get the following, does your output match? The idea here is to see that the update function is the desired one. So useful tools to investigate that are: 1. Print out the source of update and see that it is a standard generic function. 2. Use showMethods(update) and verify that there is a method listed for ProgressBarText 3. Look at find("update") Here are some examples. Perhaps if you debug (as I've done below) mas5 and step through, you will find the place for update becomes the wrong update. But I'm currently out of ideas on what is different about your setup from mine. Perhaps you can also help us there. + seth > update function (object, ...) UseMethod("update") <environment: namespace:stats=""> > library(affy) Loading required package: Biobase Loading required package: tools Welcome to Bioconductor Vignettes contain introductory material. To view, type 'openVignette()'. To cite Bioconductor, see 'citation("Biobase")' and for packages 'citation(pkgname)'. Loading required package: affyio > update standardGeneric for "update" defined from package "stats" function (object, ...) standardGeneric("update") <environment: 0x16684670=""> Methods may be defined for arguments: object > debug(mas5) > example(mas5) mas5> data(affybatch.example) mas5> eset <- mas5(affybatch.example) debugging in: mas5(affybatch.example) debug: { res <- expresso(object, bgcorrect.method = "mas", pmcorrect.method = "mas", normalize = FALSE, summary.method = "mas", ...) if (normalize) res <- affy.scalevalue.exprSet(res, sc = sc, analysis = analysis) return(res) } Browse[1]> update standardGeneric for "update" defined from package "stats" function (object, ...) standardGeneric("update") <environment: 0x16684670=""> Methods may be defined for arguments: object Browse[1]> showMethods(update) Function: update (package stats) object="ANY" object="ProgressBarText" Browse[1]> -- Seth Falcon | Computational Biology | Fred Hutchinson Cancer Research Center http://bioconductor.org