Dear Gordon Thank you for your help. Following the 3.3.2 section of the user guide: #set the references targets$Age <- relevel(targets$Age, ref="Young") targets$Genotype <- relevel(targets$Genotype, ref="Wt") #set design1 design1 <- model.matrix(~ Genotype + Genotype:Age, data=targets) > colnames(design1) [1] "(Intercept)"           "GenotypeMutant" "GenotypeWt:AgeOld"     "GenotypeMutant:AgeOld" #interpretation "GenotypeWt:AgeOld" (coefficient 3) represents the genes that change with age in Wt. "GenotypeMutant:AgeOld" (coefficient 4) represents the genes that change with age in Mutant. # To test for interaction, I thought I should follow section 3.3.4 of the user guide: design2 <- model.matrix(~ Genotype * Age, data=targets) > colnames(design2) [1] "(Intercept)"           "GenotypeMutant"        "AgeOld" "GenotypeMutant:AgeOld" "GenotypeMutant:AgeOld" (coefficient 4) represents represents the interaction of Genotype and Age I am doing this right? Thanks Joseph On Sunday, January 12, 2014 3:24 PM, Gordon K Smyth <smyth@wehi.edu.au> wrote: Dear Joseph, > >You have not fitted the right model.  You need > >    ~ Genotype + Genotype:Age > >instead of > >    ~ Age + Age:Genotype > >You can follow the section 3.3.2 in the edgeR User's Guide. > >Best wishes >Gordon > >> Date: Sat, 11 Jan 2014 12:36:07 -0800 (PST) >> From: joseph <jdsandjd@yahoo.com> >> To: "bioconductor@r-project.org" <bioconductor@r-project.org> >> Subject: [BioC] EdgeR design question? >> >> Hi >> I am using edgeR to analyse an RNA-seq experiment with two factors, age (Young, Old) and genotype (Wt, Mutant). >>> targets >>            Age Genotype >> Sample1  Young       Wt >> Sample2  Young       Wt >> Sample3  Young       Wt >> Sample4  Young   Mutant >> Sample5  Young   Mutant >> Sample6  Young   Mutant >> Sample7    Old       Wt >> Sample8    Old       Wt >> Sample9    Old       Wt >> Sample10   Old   Mutant >> Sample11   Old   Mutant >> Sample12   Old   Mutant >> >> I know from previous experiments that the age effect is different between Wt and Mutant genotypes. >> I would like to find: >> 1. genes that change with age in Wt. >> 2. genes that change with age in Mutant. >> 3. genes that change with age differently between the genotypes (interaction). >> >> Can the following approach answer these questions? >> >> targets$Age <- relevel(targets$Age, ref="Young") >> targets$Genotype <- relevel(targets$Genotype, ref="Wt") >> design1 <- model.matrix(~Age + Age:Genotype, data=targets) >>> design1 >>          (Intercept) AgeOld AgeYoung:GenotypeMutant AgeOld:GenotypeMutant >> Sample1            1      0                       0 0 >> Sample2            1      0                       0 0 >> Sample3            1      0                       0 0 >> Sample4            1      0                       1 0 >> Sample5            1      0                       1 0 >> Sample6            1      0                       1 0 >> Sample7            1      1                       0 0 >> Sample8            1      1                       0 0 >> Sample9            1      1                       0 0 >> Sample10           1      1                       0 1 >> Sample11           1      1                       0 1 >> Sample12           1      1                       0 1 >> attr(,"assign") >> [1] 0 1 2 2 >> attr(,"contrasts") >> attr(,"contrasts")$Age >> [1] "contr.treatment" >> attr(,"contrasts")$Genotype >> [1] "contr.treatment" >> >> >> I think coef=4 in design1 represents the genes that change with age in Mutant; but I don't know what coefficient represents the genes that change with age in Wt. >> >> design2 <- model.matrix(~Age*Genotype, data=targets) >>> design2 >>          (Intercept) AgeOld GenotypeMutant AgeOld:GenotypeMutant >> Sample1            1      0              0                     0 >> Sample2            1      0              0                     0 >> Sample3            1      0              0                     0 >> Sample4            1      0              1                     0 >> Sample5            1      0              1                     0 >> Sample6            1      0              1                     0 >> Sample7            1      1              0                     0 >> Sample8            1      1              0                     0 >> Sample9            1      1              0                     0 >> Sample10           1      1              1                     1 >> Sample11           1      1              1                     1 >> Sample12           1      1              1                     1 >> attr(,"assign") >> [1] 0 1 2 3 >> attr(,"contrasts") >> attr(,"contrasts")$Age >> [1] "contr.treatment" >> attr(,"contrasts")$Genotype >> [1] "contr.treatment" >> >> coef=4 in design2 represents the interaction of age and genotype. >> >> Thank you for your help >> Joseph >> >> >>> sessionInfo() >> R version 3.0.2 Patched (2013-10-28 r64119) >> Platform: x86_64-apple-darwin10.8.0 (64-bit) >> >> locale: >> [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8 >> >> attached base packages: >> [1] stats     graphics  grDevices utils     datasets  methods base >> >> other attached packages: >> [1] edgeR_3.4.2  limma_3.18.7 >> >> loaded via a namespace (and not attached): >> [1] tools_3.0.2 > >_____________________________________________________________________ _ >The information in this email is confidential and inten...{{dropped:10}}