I was reading blog posts, manuals and tutorials to find a good DEG pipeline. I thought edgeR was a good choice and tximport was good to summarise my transcript level abundances, but then I saw this tweet from a well-known PI who claims that this was a terrible choice. Is edgeR + tximport not reliable?

Link to tweet

We published a systematic assessment of DE methods downstream of tximport, in particular across a variety of sample sizes, and found edgeR to be one of the best performing methods. Click through on the image to the DGE section.