DESeq2 with multiple conditional data

Question

[Omics data mining]

Hello everyone I am mainly interested to identify the differential based on the "condition" together with "location" specific effect. Here is my few lines of metadata :

id  "condition" "location"
1   "P" "V1"
2   "P" "V1"
11  "P" "V1"
12  "P" "V1"
13  "P" "V1"
14  "P" "V1"
15  "P" CR
16  "P" CR
17  "P" CR
23  "P" CR
24  "P" CR
25  "CT"    "V1"
26  "CT"    "V1"
27  "CT"    "V1"
28  "CT"    "V1"
29  "CT"    "V1"
30  "CT"    "V1"
31  "CT"    "V1"
32  "CT"    "V1"
33  "CT"    CR
34  "CT"    CR
35  "CT"    CR
36  "CT"    CR

I used following design for DESeq2 :

design = ~ location + condition +  location:condition

From DESeq2, it gives me output for the three comparison : **"condition_P_vs_CT

location_V1_vs_CR

locationV1.conditionP**

Result from condition_P_vs_CT will give differential genes for the "condition" specific difference Result from location_V1_vs_CR will give differential genes for the "location specific difference How can I estimate the differential genes showing the "join" effect of location with condition from DESeq2.

Thanks in advance

I will appreciate all the suggestions.

[rpolicastro]

In regression when you add an interaction term it changes the meaning of the main effects. For DESeq2 in particular the main effects for (e.g.) condition becomes the effect of condition at the reference level location.

To make initial analysis easier I would first do a likelihood ratio test of the interaction term to check whether the effect of condition on a gene changes based on the location for any genes. Next I would collapse your factor levels into P_V1, P_VR, etc. so you end with with the regression formula ~ location_condition and then simply doing your contrasts of interest using those combined factor levels.