Questions about AMOVA analysis

Theory behind mothur
1

Hi everyone, i performed a 16s rRNA analysis for studying the intestinal microbiome of mice and assess the effect of chronic and metabolic stress in mices. Also i’ taking in consideration the sex.
I have 17 samples:

  • Control Group: 3 males(CM) and 3 females (CH)
    *Group expossed to chronic stress: 3 males (EM) and 3 females (EH)
    *Group for Metabolic stress: 3 males (DM) and 2 females (DH)

And, for assessing effect of the treatments, a teacher suggest me for an ANOVA analysis.

I followed the MiSeq SOP. So i got my distance matrix from the follow code :
mothur > dist.shared(shared=final.opti_mcc.shared)

My design file are in the format .tsv.

It was correct to use this files?

this is what i obtained comparing treatments (control x Stress x Metabolic stress) :

Ctl-Dta-Est Among Within Total
SS 0.892772 5.3959 6.28867
df 2 14 16
MS 0.446386 0.385421

Fs: 1.15818
p-value: 0.003*

Ctl-Dta Among Within Total
SS 0.463626 3.49937 3.96299
df 1 9 10
MS 0.463626 0.388819

Fs: 1.1924
p-value: 0.022

Ctl-Est Among Within Total
SS 0.460655 3.87668 4.33733
df 1 10 11
MS 0.460655 0.387668

Fs: 1.18827
p-value: 0.025

Dta-Est Among Within Total
SS 0.413579 3.41576 3.82934
df 1 9 10
MS 0.413579 0.379528

Fs: 1.08972
p-value: 0.096

What does that abbreviations means… SS, df , MS and Fs ? These results are enough to prove a hypothesis?

I’ve seen from another methods (for example, Rstudio) that i need a Fstat file, or a genind file with my samples information. It is posible to obtain these files from Mothur?

Thanks!

2

Hi there,

Your design is considerably more complicated than the single classification ANOVA we can do with amova in mothur. I’d strongly encourage you to check out adonis2 in the vegan R package. You can use a distance matrix as input to adonis2. You can see a video I made about using adonis (basically the same interface as adonis2) here: https://youtu.be/oLf0EpMJ4yA

In our case, SS and MS are the sum squares and mean sum squares, df is the degrees of freedom, and Fs is the F statistic.

Pat

3

Thanks you very much for your help!

I made a mistake writing … sorry :sweat_smile: i mean AMOVA. What i made in Mothur was the AMOVA analysis

But now, i’ve performer your tutorial and have mi results …that’s PerMANOVA, right?
also perform a pairwise analysis with “pairwise.adonis2”

what is the difference between this two methods?
Again, thank you for your help :smiley:

4

It’s the same basic method but adonis2 allows you to use much more sophisticated experimental designs.

Pat