Entering edit mode
Suppose that for three replicates of treatment 1, the average TPM for a particular gene is 80 (and not much difference between the three replicates).
And for three replicates of treatment 2, the average TPM for the same gene is 990 (again, not much difference between replicates).
80 / 990 = 0.08 so doesn't that mean that the Log2 fold change should be approximately -3.64 in treatment 1? I am getting a Log2 fold change of only -0.75, which seems way way off.
What's going on here?
Please add code and the actual counts (not the TPMs, the normalized counts from DESeq2) and a section of your results object for people to comment. Textual descriptions are hard to debug.