Hi,

I have some DNA methylation data of patients and their families, which are confounded by age. Thus, I'd like to adjust the model with age as a covariate to see the disease effect not influenced by age. I used the following design matrix: design<-model.matrix(~ factor+age) Is this a correct design matrix format? The rest of the analysis code bellow. Cheers! Matt

factor           <- factor(anno$Subject_group) 
age <- as.numeric(anno$Epigenetic_clock_Horvath)

design<-model.matrix(~ factor+age) 
colnames(design) <- c("Patient", "Parent", "Sibling", "Age")


fit             <- lmFit(mval, design)

contrast_matrix <- makeContrasts(Patient - Parent,
                                 Patient - Sibling,
                                 Parent - Sibling,
                                 levels = design)

fit_contrast <- contrasts.fit(fit, contrast_matrix) 
fit_ebayes   <- eBayes(fit_contrast)

Yes, that is the standard way to add age as a covariate.