I realize this question has probably been asked a million times but I cannot find a case that's similar to mine.

I have 8 conditions but no reference. These are different strains of the same organism. I want to perform a pairwise comparison of all 8 groups , totaling 28 unique comparisons (no A vs. A, but A vs. B, A vs. C, etc..)

Is this possible?

> coldata
          strain
A_S15     A
A_S23     A
A_S7      A
B_S12     B
B_S20     B
B_S4      B
C_S16     C
C_S24     C
C_S8      C
D_S1      D
D_S17     D
D_S9      D
E_S13     E
E_S21     E
E_S5      E
F_S14     F
F_S22     F
F_S6      F
G_S10     G
G_S18     G
G_S2      G
H_S11     H
H_S19     H
H_S3      H

> dds <- DESeqDataSetFromMatrix(countData=counts,colData=coldata, design= ~ strain)
> dds <- DESeq(dds)
> resultsNames(dds)
[1] "Intercept"          "strain_B_vs_A"  "strain_C_vs_A"
[4] "strain_D_vs_A" "strain_E_vs_A"  "strain_F_vs_A"
[7] "strain_G_vs_A"  "strain_H_vs_A"

So to achieve my 28 comparisons, should I re-reference the strain like this?

> dds$condition <- relevel(dds$condition, ref="B")

And are "strainBvsA" and "strainAvsB" equivalent or does the order matter?

Thank you.

No, you do not have to re-level the factor. You can easily conduct pairwise comparisons as exemplified here:

• Question: DESeq2 compare all levels
• https://support.bioconductor.org/p/60032/#60055
• Question: Deseq2 Pairwise Comparison Between Conditions

Kevin