Dear Tiandao, There is no good way to plot residuals for microarray experiments that I know of, so I am not at all sure what you have in mind here. You cannot sensibly do one big plot of all the residuals because the data is so heteroscedastic across genes, as well as being dependent. The residuals() method for MArrayLM objects doesn't work when ndups>1. Do you want residuals before or after averaging over dups? If the later, you might try MAav <- avedups(MA, ndups=4) fitav <- lmFit(MAav, design) res <- residuals(fitav, MAav) Best wishes Gordon >Date: Sat, 13 Oct 2007 14:55:24 -0500 (CDT) >From: Tiandao Li <tiandao.li at="" usm.edu=""> >Subject: [BioC] residuals.MArrayLM >To: bioconductor at stat.math.ethz.ch >Message-ID: <pine.lnx.4.64.0710131419210.30552 at="" orca.st.usm.edu=""> >Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed > >Dear List, > >I am using limma to analyze 2-color microarray data. After reading data >from gpr files, normalize with print-tip loess, then build linear model to >find the differentially expressed gene list. However, before jump to any >conclusions, I want to check the mode first, whether the linear model is a >good fit of the real data. The common check is residual plot to valid the >model assumption. However, one residuals from stats package and one >residuals.MArrayLM from limma package, basically the usages are kind of >similar, residuals(object,y,..). Now I would like to extract residuals >from model. I searched the BioC and R archives, and also google the web, I >knew how to extract residuals using lm from R, from there I knew how to >work around to extract residuals from limma model. However, there must be >a easy way to exract the residuals from limma linear model. > >Please forgive my simple question, any comments are welcome! > >Best wishes, > >Tiandao > >MA <- normalizeWithinArrays(RG) ># correlation between duplicates >design <- modelMatrix(targets,ref="REF") >corfit <- duplicateCorrelation(MA,design,ndups=4) >fit <- lmFit(MA,design,ndups=4,correlation=corfit$consensus,method="ls") > > > sessionInfo() >R version 2.5.1 (2007-06-27) >i386-pc-mingw32 > >locale: >LC_COLLATE=English_United States.1252;LC_CTYPE=English_United >States.1252;LC_MONETARY=English_United >States.1252;LC_NUMERIC=C;LC_TIME=English_United States.1252 > >attached base packages: >[1] "stats" "graphics" "grDevices" "utils" "datasets" "methods" >[7] "base" > >other attached packages: > MASS statmod limma >"7.2-34" "1.3.0" "2.10.5"

Dear Dr. Symth, Thank you so much for your valuable help. This is what I was thinking of. May I ask one more question since I searched for days, how do we objectively assess lmFit result(s) using least squares or robust methods? Regards, Tiandao On Mon, 15 Oct 2007, Gordon Smyth wrote: Dear Tiandao, There is no good way to plot residuals for microarray experiments that I know of, so I am not at all sure what you have in mind here. You cannot sensibly do one big plot of all the residuals because the data is so heteroscedastic across genes, as well as being dependent. The residuals() method for MArrayLM objects doesn't work when ndups>1. Do you want residuals before or after averaging over dups? If the later, you might try MAav <- avedups(MA, ndups=4) fitav <- lmFit(MAav, design) res <- residuals(fitav, MAav) Best wishes Gordon >Date: Sat, 13 Oct 2007 14:55:24 -0500 (CDT) >From: Tiandao Li <tiandao.li at="" usm.edu=""> >Subject: [BioC] residuals.MArrayLM >To: bioconductor at stat.math.ethz.ch >Message-ID: <pine.lnx.4.64.0710131419210.30552 at="" orca.st.usm.edu=""> >Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed > >Dear List, > >I am using limma to analyze 2-color microarray data. After reading data >from gpr files, normalize with print-tip loess, then build linear model to >find the differentially expressed gene list. However, before jump to any >conclusions, I want to check the mode first, whether the linear model is a >good fit of the real data. The common check is residual plot to valid the >model assumption. However, one residuals from stats package and one >residuals.MArrayLM from limma package, basically the usages are kind of >similar, residuals(object,y,..). Now I would like to extract residuals >from model. I searched the BioC and R archives, and also google the web, I >knew how to extract residuals using lm from R, from there I knew how to >work around to extract residuals from limma model. However, there must be >a easy way to exract the residuals from limma linear model. > >Please forgive my simple question, any comments are welcome! > >Best wishes, > >Tiandao > >MA <- normalizeWithinArrays(RG) ># correlation between duplicates >design <- modelMatrix(targets,ref="REF") >corfit <- duplicateCorrelation(MA,design,ndups=4) >fit <- lmFit(MA,design,ndups=4,correlation=corfit$consensus,method="ls") > > > sessionInfo() >R version 2.5.1 (2007-06-27) >i386-pc-mingw32 > >locale: >LC_COLLATE=English_United States.1252;LC_CTYPE=English_United >States.1252;LC_MONETARY=English_United >States.1252;LC_NUMERIC=C;LC_TIME=English_United States.1252 > >attached base packages: >[1] "stats" "graphics" "grDevices" "utils" "datasets" "methods" >[7] "base" > >other attached packages: > MASS statmod limma >"7.2-34" "1.3.0" "2.10.5" _______________________________________________ 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