Correction for multiple testing in amova and indic. sp.

Theory behind mothur
1

Hey Mothur Team,
I am woundering if the p-values that I get with the amova and the indicator species command are the uncorrected p-values or if they are already corrected for multiple testing?

Thanks for your help.
Frog

2

For AMOVA, we apply a Bonferroni correction to the experiment wise error to assess whether a p-value is significant. For indicator species, you’re on you own. But a conservative approach would be to divide 0.05 by the number of OTUs you are analyzing.

3

Hello,

You could also note that the original Bonferroni-correction is quite conservative when it comes to really many multiple comparisons. It would be also ok to use some adjusted Bonferroni methods such as Hochberg-Bonferroni or Holm-Bonferroni which are less conservative (more significant differences to be expected).

Regards, Antti Rissanen