The simplest workflow is just two-timepoint log fold change, and we should have three versions of this:

• Very simple single-replicate workflow from direct sequencing data that has a minimal amount of extra steps
• Barcoded workflow with multiple replicates (more realistic) that combines replicates using the mean and also reports a standard deviation
• Barcoded workflow with multiple replicates that uses the Enrich2 approach (natural log, replicates combined using random effects model as described in WLS regression workflow #51 )

This is still the most commonly-used way of scoring a DMS so we should use this as a training example as well.

For the example data for the latter two, we can use the https://github.com/FowlerLab/Enrich2-Example dataset, just with two time points (first/last) instead of all of them.