Amova & nan values


I’ve run the amova script on a jaccard distance metric (of otus cluster by a 3% similarity) consisting of only 4 groups (see below)

Cr 0.764732
Da 1.000000 0.860570
Wa 0.896815 0.814033 0.752521

… and Im getting a significant output, but the there are nan values as seen below. I guess you can’t use this test on such a small matrix where the groups only occur once?

Da-Wa Among Within Total
SS 0.389273 0 0.389273
df 1 0 1
MS 0.389273 nan

Fs: nan
p-value: <0.001*

This won’t work - you don’t have any replicates. You’re essentially trying to do analysis of variance with N=1