Hi ,
I was wondering if design formulas like ~ genotype + light + genotype:light can be put in current galaxy deseq2 wrapper. Also in R one can just do a `~1 ` To get just the pca plot of all samples together.
These two things would be great for recommending galaxy at my workspace.
Hi @microfuge
I created an issue here DESeq2: add support for interaction effects · Issue #8907 · galaxyproject/tools-iuc · GitHub
But the implementation might take some time. If you have a colleague who can implement it, please share the following repo to contribute. We are happy to review 
Thank you for understanding!
Pavan
Meanwhile, please try mimicking your formula with some changes to the input for the current DESeq2 wrapper.
- Use the input mode “Select contrast from sample sheet columns”
- Have your count files in a collection. Name each of the samples.
- Prepare a sample sheet file. In addition to light and genotype, add a column that defines the interaction. Something like the following
| sample |
genotype |
light |
interaction |
| WT_dark_1 |
WT |
dark |
no |
| WT_dark_2 |
WT |
dark |
no |
| WT_light_1 |
WT |
light |
no |
| WT_light_2 |
WT |
light |
no |
| mutant_dark_1 |
mutant |
dark |
no |
| mutant_dark_2 |
mutant |
dark |
no |
| mutant_light_1 |
mutant |
light |
yes |
| mutant_light_2 |
mutant |
light |
yes |
Then use the formula ~ genotype + light + interaction. I haven’t tested. That might work.