Not getting pairwise comparisons AMOVA

[Johannes]

Hi,

I have a matrix of 577 rows/columns which represent samples from different months, a whole year… below is samples for two groups in January.
I run the AMOVA, but I don’t get any pairwise comparisons between month, just one comparison for all months (as seen below). How is that? I have run similar data containing, e.g. 6 different treatments and gotten pairwise comparisons…

• 1Ao10 Jan
  1Ao11 Jan
  1Ao12 Jan
  1Ao13 Jan
  1Ao14 Jan
  1Ao15 Jan
  1Ao16 Jan
  1Ao17 Jan
  1Cr30 Jan
  1Cr31 Jan
  1Cr32 Jan
  1Cr33 Jan
  1Cr34 Jan
  1Cr35 Jan
  1Cr36 Jan
  1Cr37 Jan
  1Cr38 Jan
  1Cr39 Jan

Apr-Aug-Dec-Feb-Jan-Jul-Jun-Mar-May-Nov-Oct-Sep 
        Among Within Total
SS 0.0541001 13.7769 13.831
df 11 565 576
MS 0.00491819 0.024384

Fs: 0.201698
p-value: 1

Thanks

[westcott]

Mothur only calculates the pairwise comparisons if the p-value <= Experiment-wise error rate. In version 1.28.0 you can select sets and force the pairwise comparison if you would like.

mothur > amova(phylip=yourPhylipFile, design=yourDesignFile, sets=Jan-Feb)

would result in the pairwise comparison of Jan and Feb.

[pschloss]

I have a matrix of 577 rows/columns which represent samples from different months, a whole year… below is samples for two groups in January.
I run the AMOVA, but I don’t get any pairwise comparisons between month, just one comparison for all months (as seen below). How is that?

Well, we’re protecting you from yourself :). If there is not a significant effect at the experiment level, there won’t be one at the group by group level.

[Johannes]

Ok, thanks!