DESeq2 Multiple Contrast Correction

[jsage8]

Should an additional p-value correction be made for running multiple contrasts? Say for example I have 12 samples and with one factor that has 4 levels, A, B, C, and D.

No code has to be inserted here.If I run DESeq using Wald tests to determine the differentially expressed genes and then compare each level to the others using contrasts like this:

dds <- DESeqDataSetFromMatrix(countData = countData,
        colData = colData,
        design = ~ condition)

dds <- DESeq(dds)

results(dds, contrast=c("condition","A","B"))
results(dds, contrast=c("condition","A","C"))
results(dds, contrast=c("condition","A","D"))
results(dds, contrast=c("condition","B","C"))
results(dds, contrast=c("condition","B","D"))
results(dds, contrast=c("condition","C","D"))

Does running multiple contrasts require an additional correction to the adjusted p-value reported in the results object?

[Michael Love]

We only provide automatic adjustment in the gene dimension. The decideTests function in limma does provide functionality for comparing across contrasts. when I see what appears to me to be opportunistic pair-wise comparisons, I just try to discourage it in favor of EDA: e.g. progammatically testing hundreds of pairs instead of having a preset hypothesis.