Question about model design in for RNA-seq experiment

Question

[Mozart]

Hello there,

I was just wondering how to correctly deal with the following model design. I want to compare two different cell populations, each taken from the same mouse; 5 replicates each. RNA was extracted in two separate days.

so I have: 2 conditions (population 1 and 2), 5 pairs and 2 batches (day 1 and 2)

should I simply do the following: ~pairs+batches+condition

Am I right?

Answer 1

[ATpoint]

This looks like:

  Sample          Condition
  Condition1_rep1 ConditionA
  Condition1_rep2 ConditionA
  Condition1_rep3 ConditionA
  Condition1_rep4 ConditionA
  Condition1_rep5 ConditionA
  Condition2_rep1 ConditionB
  Condition2_rep2 ConditionB
  Condition2_rep3 ConditionB
  Condition2_rep4 ConditionB
  Condition2_rep5 ConditionB

If so, the design is ~ Condition and the above table would be what DESeq2 accepts as colData. This is sufficient to indicate the samples that belong to one group. Group membership is not indicated via the design formula as you did with pairs. If you want to include batch then add this as an additional column and use ~ batch + condition. Check by PCA first if you see signs of batch effect. If not leave it out. I hope you did not extract all of ConditionA on day1 and B on days?

[Mozart]

Thanks a lot for these information. No, the extraction was 'balanced', meaning that I extracted the first three replicates (A and B, then) first and then the remaining ones on the following day. It has been asked to do a paired analysis given the fact cells from each replicates are coming from the same mouse. But what does it mean? How can one do that on DESeq2? I think one should just follow the normal design as usual and that's it, right?

thanks

[ATpoint]

Can you post such a table as above including all the factors that are relevant such as which animal etc.?

[Mozart]

That is basically the same as you posted above. Data retrieved from mice.

  Sample          Condition
  Condition1_rep1 ConditionA
  Condition1_rep2 ConditionA
  Condition1_rep3 ConditionA
  Condition1_rep4 ConditionA
  Condition1_rep5 ConditionA
  Condition2_rep1 ConditionB
  Condition2_rep2 ConditionB
  Condition2_rep3 ConditionB
  Condition2_rep4 ConditionB
  Condition2_rep5 ConditionB