What eventually was the correct design matrix for this dataset? Peter On Mar 17, 2004, at 10:48 AM, Jordi Altirriba Guti?rrez wrote: > Thank you very much Gordon for your quick answer! > My phenoData is: >> pData(eset) > DIABETES TREATMENT > DNT1 TRUE FALSE > DNT2 TRUE FALSE > DNT3 TRUE FALSE > DT1 TRUE TRUE > DT2 TRUE TRUE > DT3 TRUE TRUE > SNT1 FALSE FALSE > SNT2 FALSE FALSE > SNT3 FALSE FALSE > ST1 FALSE TRUE > ST2 FALSE TRUE > ST3 FALSE TRUE > > (DNT=Diabetic untreated, DT=Diabetic treated, SNT=Health treated, > ST=Health untreated) > > I want to know the genes characteristics of the diabetes, the > treatment and the treatment + diabetes. Moreover when I analyse my > data with SAM and I compare Health treated vs the Health untreated I > don't see many differences, but when I compare the Diabetic treated vs > the Diabetic treated I see a lot of differences, so is correct to > apply a 2 x 2 factorial design? > Is LIMMA the correct tool to answer my questions? If it is the correct > tool, how can I do a factorial design matrix (if to do a factorial > design is correct)? (Robert Gentleman has suggested me to use the > factDesign). > Thank you very much for your time, patience and your suggestions. > Yours sincerely, > > >> From: Gordon Smyth <smyth@wehi.edu.au> >> To: Jordi Altirriba Guti?rrez <altirriba@hotmail.com> >> CC: bioconductor@stat.math.ethz.ch >> Subject: Re: [BioC] Multifactorial analysis with RMA and LIMMA of >> Affymetrix microarrays >> Date: Wed, 17 Mar 2004 11:32:16 +1100 >> >> At 07:55 AM 17/03/2004, Jordi Altirriba Guti?rrez wrote: >>> (Sorry, but I've had some problems with the HTML) >>> Hello all! >>> I am a beginner user of R and Bioconductor, sorry if my questions >>> have already been discussed previously. >>> I am studying the effects of a new hypoglycaemic drug for the >>> treatment of diabetes and I have done this classical study: >>> 4 different groups: >>> 1.- Healthy untreated >>> 2.- Healthy treated >>> 3.- Diabetic untreated >>> 4.- Diabetic treated >>> With 3 biological replicates of each group, therefore I have done 12 >>> arrays (Affymetrix). >>> I have treated the raw data with the package RMA of Bioconductor >>> according to the article ?Exploration, normalization and summaries >>> of high density oligonucleotide array probe level data? >>> (Background=RMA, Normalization=quantiles, PM=PMonly, >>> Summarization=medianpolish). >>> I am currently trying to analyse the object eset with the package >>> LIMMA of Bioconductor. I want to know what genes are differentially >>> expressed due to diabetes, to the treatment and to the combination >>> of both (diabetes + treatment), being therefore an statistic >>> analysis similar to a two-ways ANOVA). >>> So, my questions are: >>> 1.- I have created a PhenoData in RMA, will the covariates of the >>> PhenoData have any influence in the analysis of LIMMA? >> >> Not automatically. >> >>> 2.- Are these commands correct to get these results? (see below) In >>> the command TopTable, the output of coef=1 are the genes >>> characteristics of diabetes? >> >> No. >> >>> 3.- If I do not see any effect of the treatment in the healthy >>> untreated rats should I design the matrix differently? Something >>> similar to a one-way-ANOVA, considering differently the four groups: >>> ( > design<-model.matrix(~ -1+factor(c(1,1,1,2,2,2,3,3,3,4,4,4))) ). >> >> This design matrix would be very much better, i.e., it would be a >> correct matrix. You could then use contrasts to test for differences >> and interaction terms between your four groups, and that would do the >> job. >> >> If you tell us what's in your phenoData slot, i.e., type pData(eset), >> then we might be able to suggest another approach analogous to the >> classical two-way anova approach. >> >> Gordon >> >>> 5.- Any other idea? >>> Thank you very much for your time and your suggestions. >>> Yours sincerely, >>> >>> Jordi Altirriba (PhD student, Hospital Cl?nic ? IDIBAPS, Barcelona, >>> Spain) >>> >>>> design<- >>>> cbind("disease"=c(1,1,1,1,1,1,0,0,0,0,0,0),"treatment"=c(0,0,0,1,1,1 >>>> ,0,0,0,1,1,1)) >>>> fit<-lmFit(eset,design) >>>> contrast.matrix<-cbind("diabetes"=c(1,0),"drug"=c(0,1),"diabetes- >>>> drug"=c(1,1)) >>>> rownames(contrast.matrix)<-colnames(design) >>>> design >>> disease treatment >>> [1,] 1 0 >>> [2,] 1 0 >>> [3,] 1 0 >>> [4,] 1 1 >>> [5,] 1 1 >>> [6,] 1 1 >>> [7,] 0 0 >>> [8,] 0 0 >>> [9,] 0 0 >>> [10,] 0 1 >>> [11,] 0 1 >>> [12,] 0 1 >>>> contrast.matrix >>> diabetes drug diabetes-drug >>> diabetes 1 0 1 >>> tratamiento 0 1 1 >>>> fit2<-contrasts.fit(fit,contrast.matrix) >>>> fit2<-eBayes(fit2) >>>> clas<-classifyTests(fit2) >>>> vennDiagram(clas) >>>> topTable(fit2,number=100,genelist=geneNames(eset),coef=1,adjust="fdr >>>> ") >>> Name M t P.Value B >>> 11590 1387471_at 19.32575 9.442926 2.743122e-17 29.95299 >>> 2500 1369951_at 19.24748 9.404683 2.743122e-17 29.66737 >>> 10652 1384778_at 19.22968 9.395981 2.743122e-17 29.60254 >>> ? >>>> sink("limma-diabetes.txt") >>>> topTable(fit2,number=1000,genelist=geneNames(eset),coef=1,adjust="fd >>>> r") >>>> sink() >> > > _________________________________________________________________ > Reparaciones, servicios a domicilio, empresas, profesionales... Todo > en la gu?a telef?nica de QDQ. http://qdq.msn.es/msn.cfm > > _______________________________________________ > Bioconductor mailing list > Bioconductor@stat.math.ethz.ch > https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor >

Here is the answer HTH Jordi BioC] Multifactorial analysis with RMA and LIMMA of Affymetrix microarrays New Search: From: Gordon Smyth (smyth@landfield.com) Date: Thu Mar 18 2004 - 02:14:43 EST Next message: Rafael A. Irizarry: "[BioC] Subsetting Affybatch objects by gene list." Previous message: Helen Cattan: "[BioC] limma, background correction and duplicate spots" Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] ---------------------------------------------------------------------- ---------- At 02:48 AM 18/03/2004, Jordi Altirriba Guti?rrez wrote: >Thank you very much Gordon for your quick answer! My phenoData is: >>pData(eset) >DIABETES TREATMENT DNT1 TRUE FALSE DNT2 TRUE FALSE DNT3 TRUE FALSE DT1 TRUE >TRUE DT2 TRUE TRUE DT3 TRUE TRUE SNT1 FALSE FALSE SNT2 FALSE FALSE SNT3 >FALSE FALSE ST1 FALSE TRUE ST2 FALSE TRUE ST3 FALSE TRUE > >(DNT=Diabetic untreated, DT=Diabetic treated, SNT=Health treated, ST=Health >untreated) > >I want to know the genes characteristics of the diabetes, the treatment and >the treatment + diabetes. Moreover when I analyse my data with SAM and I >compare Health treated vs the Health untreated I don't see many >differences, but when I compare the Diabetic treated vs the Diabetic >treated I see a lot of differences, so is correct to apply a 2 x 2 >factorial design? You simply need to fit a model which contains four coefficient which distinguish your four groups. The classical 2x2 model is just one particular parametrization you can use: design <- model.matrix( ~ DIABETES*TREATMENT, data=pData(eset)) fit <- lmFit(eset, design) >Is LIMMA the correct tool to answer my questions? If it is the correct >tool, how can I do a factorial design matrix (if to do a factorial design >is correct)? (Robert Gentleman has suggested me to use the factDesign). You're just fitting a linear model, so the above calculation is exactly equivalent to what factDesign does, although probably a bit faster. I would use limma myself because it allows you go on to do empirical Bayes moderation of the residual standard deviations etc, which I think it important, but Robert may be able to make a further case for factDesign. Cheers Gordon >Thank you very much for your time, patience and your suggestions. Yours >sincerely, ---------------------------------------------------------------------- ---------- Next message: Rafael A. Irizarry: "[BioC] Subsetting Affybatch objects by gene list." Previous message: Helen Cattan: "[BioC] limma, background correction and duplicate spots" Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] ---------------------------------------------------------------------- ---------- This archive was generated by hypermail 2.0.0 : Fri Jul 16 2004 - 13:13:33 EDT >From: Peter Lee <peter.d.lee@mcgill.ca> >To: Jordi Altirriba Guti?rrez <altirriba@hotmail.com> >CC: bioconductor@stat.math.ethz.ch >Subject: Re: [BioC] Multifactorial analysis with RMA and LIMMA of >Affymetrix microarrays >Date: Thu, 5 Aug 2004 14:42:12 -0400 > >What eventually was the correct design matrix for this dataset? > >Peter > >On Mar 17, 2004, at 10:48 AM, Jordi Altirriba Guti?rrez wrote: > >>Thank you very much Gordon for your quick answer! >>My phenoData is: >>>pData(eset) >> DIABETES TREATMENT >>DNT1 TRUE FALSE >>DNT2 TRUE FALSE >>DNT3 TRUE FALSE >>DT1 TRUE TRUE >>DT2 TRUE TRUE >>DT3 TRUE TRUE >>SNT1 FALSE FALSE >>SNT2 FALSE FALSE >>SNT3 FALSE FALSE >>ST1 FALSE TRUE >>ST2 FALSE TRUE >>ST3 FALSE TRUE >> >>(DNT=Diabetic untreated, DT=Diabetic treated, SNT=Health treated, >>ST=Health untreated) >> >>I want to know the genes characteristics of the diabetes, the treatment >>and the treatment + diabetes. Moreover when I analyse my data with SAM >>and I compare Health treated vs the Health untreated I don't see many >>differences, but when I compare the Diabetic treated vs the Diabetic >>treated I see a lot of differences, so is correct to apply a 2 x 2 >>factorial design? >>Is LIMMA the correct tool to answer my questions? If it is the correct >>tool, how can I do a factorial design matrix (if to do a factorial design >>is correct)? (Robert Gentleman has suggested me to use the factDesign). >>Thank you very much for your time, patience and your suggestions. >>Yours sincerely, >> >> >>>From: Gordon Smyth <smyth@wehi.edu.au> >>>To: Jordi Altirriba Guti?rrez <altirriba@hotmail.com> >>>CC: bioconductor@stat.math.ethz.ch >>>Subject: Re: [BioC] Multifactorial analysis with RMA and LIMMA of >>>Affymetrix microarrays >>>Date: Wed, 17 Mar 2004 11:32:16 +1100 >>> >>>At 07:55 AM 17/03/2004, Jordi Altirriba Guti?rrez wrote: >>>>(Sorry, but I've had some problems with the HTML) >>>>Hello all! >>>>I am a beginner user of R and Bioconductor, sorry if my questions have >>>>already been discussed previously. >>>>I am studying the effects of a new hypoglycaemic drug for the treatment >>>>of diabetes and I have done this classical study: >>>>4 different groups: >>>>1.- Healthy untreated >>>>2.- Healthy treated >>>>3.- Diabetic untreated >>>>4.- Diabetic treated >>>>With 3 biological replicates of each group, therefore I have done 12 >>>>arrays (Affymetrix). >>>>I have treated the raw data with the package RMA of Bioconductor >>>>according to the article ?Exploration, normalization and summaries of >>>>high density oligonucleotide array probe level data? (Background=RMA, >>>>Normalization=quantiles, PM=PMonly, Summarization=medianpolish). >>>>I am currently trying to analyse the object eset with the package LIMMA >>>>of Bioconductor. I want to know what genes are differentially expressed >>>>due to diabetes, to the treatment and to the combination of both >>>>(diabetes + treatment), being therefore an statistic analysis similar >>>>to a two-ways ANOVA). >>>>So, my questions are: >>>>1.- I have created a PhenoData in RMA, will the covariates of the >>>>PhenoData have any influence in the analysis of LIMMA? >>> >>>Not automatically. >>> >>>>2.- Are these commands correct to get these results? (see below) In the >>>>command TopTable, the output of coef=1 are the genes characteristics of >>>>diabetes? >>> >>>No. >>> >>>>3.- If I do not see any effect of the treatment in the healthy >>>>untreated rats should I design the matrix differently? Something >>>>similar to a one-way-ANOVA, considering differently the four groups: >>>>( > design<-model.matrix(~ -1+factor(c(1,1,1,2,2,2,3,3,3,4,4,4))) ). >>> >>>This design matrix would be very much better, i.e., it would be a >>>correct matrix. You could then use contrasts to test for differences and >>>interaction terms between your four groups, and that would do the job. >>> >>>If you tell us what's in your phenoData slot, i.e., type pData(eset), >>>then we might be able to suggest another approach analogous to the >>>classical two-way anova approach. >>> >>>Gordon >>> >>>>5.- Any other idea? >>>>Thank you very much for your time and your suggestions. >>>>Yours sincerely, >>>> >>>>Jordi Altirriba (PhD student, Hospital Cl?nic ? IDIBAPS, Barcelona, >>>>Spain) >>>> >>>>>design<- >>>>>cbind("disease"=c(1,1,1,1,1,1,0,0,0,0,0,0),"treatment"=c(0,0,0,1, 1,1 >>>>>,0,0,0,1,1,1)) >>>>>fit<-lmFit(eset,design) >>>>>contrast.matrix<-cbind("diabetes"=c(1,0),"drug"=c(0,1),"diabetes- >>>>>drug"=c(1,1)) >>>>>rownames(contrast.matrix)<-colnames(design) >>>>>design >>>> disease treatment >>>>[1,] 1 0 >>>>[2,] 1 0 >>>>[3,] 1 0 >>>>[4,] 1 1 >>>>[5,] 1 1 >>>>[6,] 1 1 >>>>[7,] 0 0 >>>>[8,] 0 0 >>>>[9,] 0 0 >>>>[10,] 0 1 >>>>[11,] 0 1 >>>>[12,] 0 1 >>>>>contrast.matrix >>>> diabetes drug diabetes-drug >>>>diabetes 1 0 1 >>>>tratamiento 0 1 1 >>>>>fit2<-contrasts.fit(fit,contrast.matrix) >>>>>fit2<-eBayes(fit2) >>>>>clas<-classifyTests(fit2) >>>>>vennDiagram(clas) >>>>>topTable(fit2,number=100,genelist=geneNames(eset),coef=1,adjust=" fdr ") >>>> Name M t P.Value B >>>>11590 1387471_at 19.32575 9.442926 2.743122e-17 29.95299 >>>>2500 1369951_at 19.24748 9.404683 2.743122e-17 29.66737 >>>>10652 1384778_at 19.22968 9.395981 2.743122e-17 29.60254 >>>> >>>>>sink("limma-diabetes.txt") >>>>>topTable(fit2,number=1000,genelist=geneNames(eset),coef=1,adjust= "fd >>>>>r") >>>>>sink() >>> >> >>_________________________________________________________________ >>Reparaciones, servicios a domicilio, empresas, profesionales... Todo en >>la gu?a telef?nica de QDQ. http://qdq.msn.es/msn.cfm >> >>_______________________________________________ >>Bioconductor mailing list >>Bioconductor@stat.math.ethz.ch >>https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor >> > m?s opciones. http://www.msn.es/Viajes/