Beta-group-significance q-value correction ?

[Larallbb (Lara Labarta Bajo)]

Dear all, I hope everyone is staying healthy.

I have a question regarding beta_group_significance with PERMANOVA followed by pairwise comparisons. Which method does Qiime2 use to correct the p-value (which becomes a q value) for multiple comparisons? I can’t find that in the documentation of beta_group_significance or in the original paper (Anderson MJ, 2001). Does anybody know if this is an FDR correction using Bonferroni, Benjamini & Hochberg or other?

Additionally, in order to obtain R-squared values for each test I perform, do I need to run adonis? How is that different from just running beta_group_significance using PERMANOVA?

Any help is very much appreciated. Stay safe

Lara

[Nicholas_Bokulich]

Welcome to the forum @Larallbb!

@Larallbb

Which method does Qiime2 use to correct the p-value (which becomes a q value) for multiple comparisons?

Benjamini & Hochberg

@Larallbb

in order to obtain R-squared values for each test I perform, do I need to run adonis?

yes

@Larallbb

How is that different from just running beta_group_significance using PERMANOVA?

Same test, but a different implementation. beta_group_significance uses the scikitbio implementation and generates some neat visualizations but only one-way permanova. adonis uses the vegan adonis implementation in R and supports multi-way permanova but doesn't generate the neat visualizations.

I hope that helps!