Hi,

I am not a statistician so apologies in advance if my question is naive. This is a more a methodological question, than a specific issue with code. I am dealing with data from a bulk RNASEq experiment with cells grown in different environments, treated with different agents at different concentrations and in different combinations of single cell cultures, co-cultures and triple co-cultures. . My question is to do with the design formula for DeSeq2. I have a number of comparisons to make. These comparisons can be single factor and or additive between various combinations of the factors. My question is would I need to make a new design formula and re-run the Deseq2 process each time for each unique comparison, or can I code up the distinct groups as a single factor and use that as the design for every comparison that I need to make. Which of these two approaches could be considered best practise?

Thanks in advance,

Saif

Enter the body of text here

Code should be placed in three backticks as shown below

# include your problematic code here with any corresponding output 
# please also include the results of running the following in an R session 

sessionInfo( )

Usually one would use a single design and extract comparisons. With a complex design, you should partner with someone familiar with linear models. The step of formulating the statistical design and extracting results is very important, to properly interpret the experiment.