I want to run adonis2() to look for differences in my distance matrix. My samples are grouped based on their type (category of 4), which is baseline non response, post non response, baseline response and post response. I also want to see if there is any interaction considering the overall survival of the patients (<1 year, 1-2 year, >2 year).
So far I am using this design:
adonis2(bray_dist ~ sample_data(physeq)$Group*sample_data(physeq)$Overall.Survival)
Permutation test for adonis under reduced model
Terms added sequentially (first to last)
Permutation: free
Number of permutations: 999
adonis2(formula = bray_dist ~ sample_data(physeq)$Group * sample_data(physeq)$Overall.Survival)
Df SumOfSqs R2 F Pr(>F)
sample_data(physeq)$Group 3 1.2509 0.05443 0.9892 0.552
sample_data(physeq)$Overall.Survival 2 0.9077 0.03950 1.0767 0.170
sample_data(physeq)$Group:sample_data(physeq)$Overall.Survival 4 1.4333 0.06237 0.8501 0.999
Residual 46 19.3899 0.84371
Total 55 22.9818 1.00000
However I am aware that changing the order of these variables changes the result of the test:
adonis2(bray_dist ~ sample_data(physeq)$Overall.Survival*sample_data(physeq)$Group)
Permutation test for adonis under reduced model
Terms added sequentially (first to last)
Permutation: free
Number of permutations: 999
adonis2(formula = bray_dist ~ sample_data(physeq)$Overall.Survival * sample_data(physeq)$Group)
Df SumOfSqs R2 F Pr(>F)
sample_data(physeq)$Overall.Survival 2 1.0171 0.04426 1.2065 0.013 *
sample_data(physeq)$Group 3 1.1414 0.04967 0.9026 0.932
sample_data(physeq)$Overall.Survival:sample_data(physeq)$Group 4 1.4333 0.06237 0.8501 0.998
Residual 46 19.3899 0.84371
Total 55 22.9818 1.00000
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Why do the results change and what is the best way to run this test?
