I was wondering if it is possible to use the 'contrast' keyword of the 'results' function of DESeq2 in order to select genes that are specifically expressed in a single biological class in cases where we have more than 2 conditions.
In this case I would just calculate contrasts for all the condition and e.g. select genes that are 'upregulated' in a single condition vs all the rest. Would this work? Would I have to correct the results for multiple comparisons again?
e.g., in this experimental design with a factor of three values (conditions):
samples <- read.table("samples.txt",sep="\t",as.is=T) colnames(samples)<-c("sampleName","fileName","condition") dds1 <- DESeqDataSetFromHTSeqCount(sampleTable=samples,directory="./",design= ~ condition) dds1 <- DESeq(dds1) res <- results(dds1)
then I do pairwise comparison:
res.A.B <- results(dds1, contrast=c("condition","A","B")) res.A.C <- results(dds1, contrast=c("condition","A","C")) res.B.C <- results(dds1, contrast=c("condition","B","C"))
At this point can I select differentially expresses genes that are specifically expressed in e.g. 'A' only by intersecting the results? I do not need to correct for multiple comparisons? Can this same principle be extended to case where I have e.g. 10 levels to find genes that are specifically expressed in one condition (computationally it is not a great burden, since the 'parallel' option comes to the rescue)?
Thank you in advance