Hello, I have some R/BioC code that was working and is not working any more, and I suspect that it has to do with the fact that I did a global "update packages". The error comes from trying to create a pseudo image of a CEL file which is part of an Affy batch, exactly as is done in the vignette "affyPLM: Model Based QC Assessment of Affymetrix". > aBatRef <- ReadAffy() > Pset <- fitPLM(aBatRef) > image(Pset, which =2) Erreur dans image.default(Pset) : argument must be matrix-like I found a similar problem in the BioC help archives and the asker was advised to reinstall R, BioC , and all packages. So I reinstalled everything with the latest stable versions of R and BioC and I still get the error. Looking in the source code for image.R, I see that his particular error is reported when the dimensions of the objet received are null, and indeed that is the case: > dim(Pset) NULL > class(Pset) [1] "PLMset" attr(,"package") [1] "affyPLM" So I tried to use the "Dilution" data set as described in the vignette "affyPLM: Model Based QC Assessment of Affymetrix" (April 25, 2007). The only change in the last year to the vignette (with respect to creation of the pseudo image) is the use of "updateObject" library(affyPLM) data(Dilution) Dilution = updateObject(Dilution) Pset <- fitPLM(Dilution) image(Pset, which = 2) The "Dilution" data set also results in the same error, so I thought that I would report it to you ! Do you reproduce this ??? Also, in the fitPLM vignette, the examples were "not run" (I have never seen that before :) Examples data(affybatch.example) Pset <- fitPLM(affybatch.example,model=PM ~ -1 + probes + samples) se(Pset)[1:5,] # A larger example testing weight image function data(Dilution) ## Not run: Pset <- fitPLM(Dilution,model=PM ~ -1 + probes + samples) ## Not run: image(Pset) ## Not run: NUSE(Pset) # NUSE Thanks for your help. PS - I also tried this : image(Pset at residualSE). I get an image but it is definitely NOT what I was expecting! > sessionInfo() R version 2.5.1 (2007-06-27) i386-pc-mingw32 locale: LC_COLLATE=French_France.1252;LC_CTYPE=French_France.1252;LC_MONETARY= French _France.1252;LC_NUMERIC=C;LC_TIME=French_France.1252 attached base packages: [1] "grDevices" "datasets" "splines" "graphics" "stats" "utils" "tools" "methods" "base" other attached packages: hgu95av2cdf mouse4302cdf RColorBrewer affyPLM gcrma matchprobes affydata simpleaffy genefilter survival "1.16.0" "1.16.0" "0.2-3" "1.12.0" "2.8.1" "1.8.1" "1.11.2" "2.10.31" "1.14.1" "2.32" affy affyio Biobase "1.14.2" "1.4.1" "1.14.1"

Hi Rebecca, Rebecca wrote: > Hello, > I have some R/BioC code that was working and is not working any more, and I > suspect that it has to do with the fact that I did a global "update > packages". The error comes from trying to create a pseudo image of a CEL > file which is part of an Affy batch, exactly as is done in the vignette > "affyPLM: Model Based QC Assessment of Affymetrix". > > >> aBatRef <- ReadAffy() >> Pset <- fitPLM(aBatRef) >> image(Pset, which =2) >> > > Erreur dans image.default(Pset) : argument must be matrix-like > This error indicates that you are not getting the correct method for your PLMset object, which indicates some problem with your installation. If you load the affyPLM package and type showMethods("image", class="PLMset") do you get this? > showMethods("image", class="PLMset") Function: image (package graphics) x="PLMset" > > I found a similar problem in the BioC help archives and the asker was > advised to reinstall R, BioC , and all packages. > So I reinstalled everything with the latest stable versions of R and BioC > and I still get the error. > How did you re-install? Did you use biocLite() directly, or did you do an update.packages()? Best, Jim > Looking in the source code for image.R, I see that his particular error is > reported when the dimensions of the objet received are null, and indeed that > is the case: > > >> dim(Pset) >> > NULL > > >> class(Pset) >> > [1] "PLMset" > attr(,"package") > [1] "affyPLM" > > So I tried to use the "Dilution" data set as described in the vignette > "affyPLM: Model Based QC Assessment of Affymetrix" (April 25, 2007). The > only change in the last year to the vignette (with respect to creation of > the pseudo image) is the use of "updateObject" > > library(affyPLM) > data(Dilution) > Dilution = updateObject(Dilution) > Pset <- fitPLM(Dilution) > image(Pset, which = 2) > > The "Dilution" data set also results in the same error, so I thought that I > would report it to you ! Do you reproduce this ??? > > Also, in the fitPLM vignette, the examples were "not run" (I have never seen > that before :) > Examples > data(affybatch.example) > Pset <- fitPLM(affybatch.example,model=PM ~ -1 + probes + samples) > se(Pset)[1:5,] > > # A larger example testing weight image function > data(Dilution) > ## Not run: Pset <- fitPLM(Dilution,model=PM ~ -1 + probes + > samples) > ## Not run: image(Pset) > ## Not run: NUSE(Pset) # NUSE > > > Thanks for your help. > PS - I also tried this : image(Pset at residualSE). I get an image but it is > definitely NOT what I was expecting! > > > >> sessionInfo() >> > R version 2.5.1 (2007-06-27) > i386-pc-mingw32 > > locale: > LC_COLLATE=French_France.1252;LC_CTYPE=French_France.1252;LC_MONETAR Y=French > _France.1252;LC_NUMERIC=C;LC_TIME=French_France.1252 > > attached base packages: > [1] "grDevices" "datasets" "splines" "graphics" "stats" "utils" > "tools" "methods" "base" > > other attached packages: > hgu95av2cdf mouse4302cdf RColorBrewer affyPLM gcrma > matchprobes affydata simpleaffy genefilter survival > "1.16.0" "1.16.0" "0.2-3" "1.12.0" "2.8.1" > "1.8.1" "1.11.2" "2.10.31" "1.14.1" "2.32" > affy affyio Biobase > "1.14.2" "1.4.1" "1.14.1" > > _______________________________________________ > Bioconductor mailing list > Bioconductor at stat.math.ethz.ch > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor > ********************************************************** Electronic Mail is not secure, may not be read every day, and should not be used for urgent or sensitive issues.

Rebecca wrote: > Thank you James, > Yes, I did use BiocLite to install and got the same results as you for the > "showMethods". > However, after carefully retracing my steps, I now know what I did that > caused the error, but no idea why, or how to fix it. > > What I did that caused the error was that I put the following line at the > end of the file "Rprofile.site" in the "~\R-2.5.1\etc" directory > > source(file.path(R.home(), "etc", "myRcode.R")) > What is in myRcode.R? Specifically, is there anything in there that might affect the class of the PLMset you are producing? What do you get from class(Pset)? Best, Jim > Where "myRcode.R" is some code I wrote for an application in R/BioC, > including the PLMset manipulations. It is in "~\R-2.5.1\etc", next to > "Rprofile.site". > > I put that line in "Rprofile.site" because I would like to be able to > double-click on the R desktop icon and have R go to the correct directory > and source the script. In other words, I don't want to have to change the > working directory and source the code. > > Now when I double click my desk top icon, I get the desired behavior, and my > home made functions are working except the pseudo image creation from the > PLMset. As you suggest, it would seem that I am not getting the correct > method for the PLMset object. > > I noticed that there are two files called "image.R" - one in > library\graphics and the other in library\marray. It must be defaulting to > the CRAN one and not the BioC one. But in that case, why would showMethods() > give the following ? > > aBat <- ReadAffy() > > Pset <- fitPLM(aBat) > > image(Pset, which =2) > Erreur dans image.default(Pset, which = 2) : argument must be > matrix-like > > showMethods("image", class="PLMset") > Function: image (package graphics) > x="PLMset" > > If you have any advise on how I can get my application to behave correctly, > I'd really appreciate it. (Rename the BioC version of image.R to something > unique?) > > Thanks again for your kind help. > > >> sessionInfo() >> > R version 2.5.1 (2007-06-27) > i386-pc-mingw32 > > locale: > LC_COLLATE=French_France.1252;LC_CTYPE=French_France.1252;LC_MONETAR Y=French > _France.1252;LC_NUMERIC=C;LC_TIME=French_France.1252 > > attached base packages: > [1] "grDevices" "datasets" "splines" "graphics" "stats" "utils" > "tools" "methods" "base" > > other attached packages: > mouse4302cdf RColorBrewer affyPLM gcrma matchprobes > affydata simpleaffy genefilter survival affy > "1.16.0" "0.2-3" "1.12.0" "2.8.1" "1.8.1" > "1.11.2" "2.10.31" "1.14.1" "2.32" "1.14.2" > affyio Biobase > "1.4.1" "1.14.1" > > > > > ********************************************************** Electronic Mail is not secure, may not be read every day, and should not be used for urgent or sensitive issues.

Hi Rebecca, I think your problem arises because of the fact that the code in the Rprofile.site is loaded into the base package. From the 'Startup' help page: Then R searches for the site-wide startup profile unless the command line option '--no-site-file' was given. The name of this file is taken from the value of the 'R_PROFILE' environment variable. If this variable is unset, the default is '$R_HOME/etc/Rprofile.site', which is used if it exists (it contains settings from the installer in a "factory-fresh" installation). This code is loaded into package 'base'. Users need to be careful not to unintentionally overwrite objects in base, and it is normally advisable to use 'local' if code needs to be executed: see the examples. I personally use an .Rprofile file in my home directory to automatically load things I think I will be needing (e.g., source("http://www.bioconductor.org/biocLite.R"), which can get pretty tedious to type). This is probably a much safer way to go, although you could try wrapping your source() call in local() as shown in the examples. Anyway, Rprofile.site is designed to be a site-wide initialization of R, which isn't nearly as useful for Windows as compared to something like a Linux server that multiple people may be using at one time. As an aside, you can figure out what your home directory is using Sys.getenv("R_USER") Best, Jim Rebecca wrote: > Hello, thanks for your response. > > > > -----Message d'origine----- > De : James W. MacDonald [mailto:jmacdon at med.umich.edu] > Envoy? : lundi 16 juillet 2007 19:00 > ? : Rebecca > Cc : bioconductor at stat.math.ethz.ch > Objet : Re: [BioC] affyPLM - cant plot chip pseudo image > > Rebecca wrote: > >> Thank you James, >> Yes, I did use BiocLite to install and got the same results as you for the >> "showMethods". >> However, after carefully retracing my steps, I now know what I did that >> caused the error, but no idea why, or how to fix it. >> >> What I did that caused the error was that I put the following line at the >> end of the file "Rprofile.site" in the "~\R-2.5.1\etc" directory >> >> source(file.path(R.home(), "etc", "myRcode.R")) >> >> > > What is in myRcode.R? Specifically, is there anything in there that > might affect the class of the PLMset you are producing? What do you get > from class(Pset)? > > > I have whittled down the code in "myRcode.R" to one line : > "library(affyPLM)". > > If I source "myRcode.R" by hand, with no changes to "Rprofile.site", I have > no trouble making a pseudo image from my saved PLM set. > > If I add the line > source(file.path(R.home(), "etc", "myRcode.R"))) > to the bottom of "Rprofile.site", and open the R GUI from the desktop, > pseudo image no longer works. > > I include gory details of the tests (1,2, and 3) that show this down below. > > Maybe the key is in the fact that the two different ways of sourcing > "myRcode.R" print different traces to the GUI. > > #------------------------------------------ > ### WHEN I SOURCE BY HAND, I GET THE FOLLOWING > # AND PSEUDO IMAGE CREATION WORKS > >> source("myRcode.R") >> > Le chargement a n?cessit? le package : affy > Le chargement a n?cessit? le package : Biobase > Le chargement a n?cessit? le package : tools > > Welcome to Bioconductor > > Vignettes contain introductory material. To view, type > 'openVignette()'. To cite Bioconductor, see > 'citation("Biobase")' and for packages 'citation(pkgname)'. > > Le chargement a n?cessit? le package : affyio > Le chargement a n?cessit? le package : affydata > Le chargement a n?cessit? le package : gcrma > Le chargement a n?cessit? le package : matchprobes > Le chargement a n?cessit? le package : splines > > > > > #----------------------------------- > # WHEN OPENING R FROM THE DESKTOP, WITH "myRcode.R" SOURCED INSIDE > # "Rprofile.site", THE FOLLOWING MESSAGES APPEAR IN THE GUI. > # AND PSEUDO IMAGE CREATION DOES NOT WORK > > Le chargement a n?cessit? le package : Biobase > Le chargement a n?cessit? le package : tools > Le chargement a n?cessit? le package : utils > > Welcome to Bioconductor > > Vignettes contain introductory material. To view, type > 'openVignette()'. To cite Bioconductor, see > 'citation("Biobase")' and for packages 'citation(pkgname)'. > > Le chargement a n?cessit? le package : affyio > Le chargement a n?cessit? le package : affydata > Le chargement a n?cessit? le package : gcrma > Le chargement a n?cessit? le package : matchprobes > Le chargement a n?cessit? le package : splines > Le chargement a n?cessit? le package : stats > > Attachement du package : 'stats' > > > The following object(s) are masked from package:affy : > > update > > #---------------------------------- > # If I add the line "library(stats)" above "library(afftPLM)" > # in "myRcode.R", The introductory message is cleaner, > # but the pseudo image creation still fails. > > > > Le chargement a n?cessit? le package : affy > Le chargement a n?cessit? le package : Biobase > Le chargement a n?cessit? le package : tools > Le chargement a n?cessit? le package : utils > > Welcome to Bioconductor > > Vignettes contain introductory material. To view, type > 'openVignette()'. To cite Bioconductor, see > 'citation("Biobase")' and for packages 'citation(pkgname)'. > > Le chargement a n?cessit? le package : affyio > Le chargement a n?cessit? le package : affydata > Le chargement a n?cessit? le package : gcrma > Le chargement a n?cessit? le package : matchprobes > Le chargement a n?cessit? le package : splines > > >> dir() >> > [1] "Makeconf" "myRcode.R" "Pset.Rdata" "Rconsole" "Rdevga" > "repositories" > [7] "rgb.txt" "Rprofile.site" > > >> load("Pset.Rdata") # This is a saved PLMset which loads and >> > # makes pseudo images OK (see below) > >> image(Pset, which =2) >> > Erreur dans image.default(Pset, which = 2) : argument must be matrix-like > > >> showMethods("image", class="PLMset") >> > Function: image (package graphics) > x="PLMset" > > >> class(Pset) >> > [1] "PLMset" > attr(,"package") > [1] "affyPLM" > > ####################### GORY DEBUG DETAILS ############################ > I created an affy batch with 2 CEL files (aBat <- ReadAffy()). > Then I created a PLMset (Pset <- fitPLM(aBat)) and saved it for testing > (file "Pset.Rdata"). > Moved the file "Pset.Rdata" to "\R-2.5.1\etc". NO CHANGES to original > "Rprofile.site". > > 1- Open R 2.5.1 by double clicking desk top icon, go to "\R-2.5.1\etc". Load > "Pset.Rdata". > Verify that the pseudo image works. > > > dir() > [1] "Makeconf" "myRcode.R" "Pset.Rdata" "Rconsole" > [5] "Rdevga" "repositories" "rgb.txt" "Rprofile.site" > > library(affyPLM) > > load("Pset.Rdata") > > class(Pset) > [1] "PLMset" > attr(,"package") > [1] "affyPLM" > > image(Pset, which =2) # THIS WORKS > >q("no") > > > 2- Put command for loading library affyPLM in "myRcode.R", located in > "\R-2.5.1\etc". Still no changes to "Rprofile.site". > -------------------------- > ### "myRcode.R" contains this line only: > > library(affyPLM) > > --------------------------- > Open R 2.5.1 by double clicking desk top icon, go to "\R-2.5.1\etc". > Source the little script and verify that the pseudo image works. > > dir() > [1] "Makeconf" "myRcode.R" "Pset.Rdata" "Rconsole" > [5] "Rdevga" "repositories" "rgb.txt" "Rprofile.site" > > source("myRcode.R") > > > load("Pset.Rdata") > > image(Pset, which =2) # THIS WORKS > >q("no") > > > 3- Make changes to "Rprofile.site". > --------------------------- > ### "Rprofile.site" was edited to include these lines at the end > source(file.path(R.home(), "etc", "myRcode.R")) > --------------------------- > Open R 2.5.1 by double clicking desk top icon, go to "\R-2.5.1\etc". > > > dir() > [1] "Makeconf" "myRcode.R" "Pset.Rdata" "Rconsole" > [5] "Rdevga" "repositories" "rgb.txt" "Rprofile.site" > > load("Pset.Rdata") > > image(Pset, which =2) > Erreur dans image.default(Pset, which = 2) : argument must be > matrix-like > > > > > > > -- 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.