Hi! I am writing to ask advice on a step in an RNAseq analysis I am doing to compare samples from two groups: naive and vehicle. I started with 10x genomics data and used the Seurat pipeline to QC and run SCTransform. I then used the Dorothea and VIPER packages to generate a matrix of transcription factor normalized enrichment scores (NES), and am trying to understand: 1) the appropriate way to statistically compare TF scores between the groups and 2) whether the vehicle NES can be "normalized" to the naive one, presenting the data as fold change of NES between vehicle and naive.

Here is an example of the summarised NES data I am working with:

```
A tibble: 196 × 5
|tf |Treatment |meanNES |stDev | N|
|------|----------|--------|--------|---|
1 Ar naive -1.70 1.23 5
2 Ar veh -1.58 1.23 6
3 Arntl naive -0.707 1.50 5
4 Arntl veh -0.832 1.51 6
5 Atf1 naive -0.462 0.873 5
6 Atf1 veh -0.350 0.888 6
7 Atf2 naive -0.669 0.937 5
8 Atf2 veh -0.517 0.985 6
9 Atf4 naive -0.696 0.935 5
10 Atf4 veh -0.617 0.959 6
… with 186 more rows
```

1) Assuming normality, would it make sense to run a Student's T test on the NES values? 2) Would it make sense to calculate "percent change" of each TF using (veh-naive)/abs(naive)? If not, what is an alternate way to present vehicle data in relation to the naive "baseline"?

Any guidance would be appreciated!

Cheers