Skip to content

Permanova result interpretation

2 messages · Tingting Chen, Botta-Dukát Zoltán

Original
Tingting Chen
# 
Dear list,

I have did  PCoA, UPGMA cluster analysis, and adonis on my data. And I got
no clear separation on the PCoA plot and clustering. However the Permanova
test is giving a very small P value. In this case, can I say these two
groups are significantly different?

Thank you very much for any help.

Best,
Tingting


Call:
adonis(formula = mat ~ Condition, data = meta, permutations = 9999)

Permutation: free
Number of permutations: 9999

Terms added sequentially (first to last)

          Df SumsOfSqs  MeanSqs F.Model      R2 Pr(>F)
Condition  1   0.23217 0.232167  6.3284 0.14606  1e-04 ***
Residuals 37   1.35740 0.036686         0.85394
Total     38   1.58956                  1.00000
---
Signif. codes:  0 ?***? 0.001 ?**? 0.01 ?*? 0.05 ?.? 0.1 ? ? 1
--------------------

Tingting Chen

PhD candidate

Whistler Center for Carbohydrate Research

Department of Food Science

Purdue University

745 Agriculture Mall Drive

West Lafayette, IN 47907
Botta-Dukát Zoltán
# 
Dear Tingting,

If you tested an a priori grouping by adonis, you can say that these 
groups significantly different. It means that their difference is bigger 
than the random expectation (= exepected difference between random 
groups with same size).

However, PCoA and UPGMA results indicates that this is not the best 
grouping, i.e. there may be grouping where two groups are more separated.

Best wishes

Zoltan

2016.06.06. 21:32 keltez?ssel, Tingting Chen ?rta: