Which measure to use for gene Fold change analysis between pre- and post-treatment samples and for gene comparison within a sample about tximport

[mikelove]

If you have follow up questions, let's continue on the support site, as this is where most people go to look for tximport usage questions.

"the way to go is ... the estimated counts...with the edgeR pipeline described in the tximport vignette"

Yes, the default recommendation is to import estimated counts and use the gene lengths as a per-sample offset. This is what happens if you use the DESeq2 or edgeR code in the vignette.

For CPM vs TMM-scaling, you'll have to consult the edgeR vignette.

"...suggest that processing the data with the estimated ("raw") counts with the edgeR pipeline has a similar effect as using scaledTPM or lengthscaledTPM"

No, raw counts alone are not similar to scaledTPM or lengthScaledTPM. It is the count + offset that gives a similar result, not count analysis alone.

"what measure would you suggest to use for calculating logFC"

The LFC estmated by DESeq2 or edgeR and returned in the results table or topTable after using any of the tximport recommended pipelines (estimated count + offset, or the counts-from-abundance option).