Differentially expression from a TPM-normalized single-cell data frame

Question

[Palgrave]

I have downloaded an expression matrix from a single-cell experiment that is already TPM normalized. What is the correct way yo preceded to detect differentially expressed genes based on this matrix? Does it have to be normalized further or can I use it directly as input in for instance Limma?

[ATpoint]

You will find advice if using the search function, as this has really been discussed dozens of times. Either use limma-trend pipeline or something like the Wilcox test. Please read available threads on DE when starting from normalized counts, as said there are many both here and at support.bioconductor.org. Then there is also https://osca.bioconductor.org/ which is always worth a read.