Appropriate Design Matrix

[StephenRicher]

Hi,

I am working with an experimental design similar to the one below and was wondering how best to fit my full and reduced model?

Currently I have fit this:

`so <- sleuth_fit(so, ~treatment, 'reduced')`
`so <- sleuth_fit(so, ~treatment + response, 'full')`

Which seems OK, but presumably the 'treatment' replicates cannot be treated equally (i.e. treatment replicates occur both between and within groups). Do you think this will be a problem?

I thought an alternative approach would be to treat within-group treatments as technical replicates and simply merge them. However, I am not sure how I would go about merging the Kallisto quantification, or whether this can be done by Sleuth?

Thanks very much,
Stephen