Reputation: 33
I'm doing some differential expression analysis between two conditions (= Timepoint) and I'm including biological covariates in my design. I want to analyse further if differences between Timepoint are different in Men and Women.
So this is my design matrix:
Timepoint <- factor(meta.data_sub$Group3)
Age <- as.numeric(meta.data_sub$Age)
Sex <- factor(meta.data_sub$Sex2)
Ethnicity <- factor(meta.data_sub$Ethnicity)
design <- model.matrix(~Age +Ethnicity +Sex +Timepoint:Sex)
head(design)
(sorry can't figure out how to fix the formating of the dataframes)
A matrix: 6 × 17 of type dbl
(Intercept) Age EthnicityB EthnicityC EthnicityD EthnicityE EthnicityF EthnicityH EthnicityJ EthnicityK EthnicityL EthnicityM EthnicityN EthnicityR Sex2 Sex1:Timepoint1 Sex2:Timepoint1
1 1 50.5 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0
2 1 39.7 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0
3 1 33.5 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0
4 1 51.7 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0
5 1 30.5 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0
6 1 37.7 0 0 0 0 0 0 0 0 0 0 0 0 0 0 00 0 0 0 0 0
And then I do
dge <- estimateDisp(dge, design = design)
fit <- glmQLFit(dge, design = design,dispersion=dge$common.dispersion, coef=16:17)
qlf <- glmQLFTest(fit, coef=16:17)
The output looks like
tt <- topTags(qlf, n = Inf)
head(tt)
logFC.Sex1.Timepoint1 logFC.Sex2.Timepoint1 logCPM F PValue FDR
APOA1 0.35602471 0.3927851 15.964477 70.22736 5.186215e-23 1.747755e-20
Does logFC.Sex1.Timepoint1 mean fold change in sex 1 between timepoint 0 and 1? And is the PValue / FDR comparing sexes or timepoints?
Thanks!
Upvotes: 1
Views: 46