Open ewijaya opened 2 years ago
I don't follow any of the pipes, but I will say that normalizeCounts
will produce, by default, log2-normalized values. So your log-fold change calculation should not be log2(negative/positive)
, but instead, negative - positive
, assuming that both negative
and positive
represent the mean log-expression value in their respective groups.
Hi Aaron,
Thank you for making this wonderful tool.
This is not an issue but a question for
scran::pairwiseTTests().
And I also know this is a scuttle repository, but I can't find one match for scran. So I posted here.I'm wondering if there's a way we can recover the mean gene expression used for calculating the fold change of
scran::pairwiseTTests
.I tried the following code used for calculating fold change using
scran::pairwiseTTests
:It produces
Notice that the logFC for Gene_1740 is 1.20 and Gene_0691 is 0.968.
But when I compute manually to get mean expression:
I get
Notice the difference in logFC. Is there a way I can easily recover the actual expression so that the logFC match calculated with
scran::pairwiseTTests
?Thanks and hope to hear from you again.
Edward