significant test in Amova and P value

[Paty Romero]

Hi, 

I have some issues interpreting my data. I'm running AMOVA, and using the significant test (with 1000 permutation) to determine if my populations are different or not. 

From the AMOVA output, almost 94% of the variation comes from within samples, 6% from between populations.The Phi value, is really low 0.05 within samples and 0.06 between populations. I'll assume with these small Phi values that there is a low level of differentiation in my populations. When I ran the significant test, the p-value for variation within samples is 0.230, supporting the idea that there is no significant population differentiation at this stratification, however, the p value between population is 0.007. 

Could be possible to say that we observed significant differentiation between my populations with a p-value of 0.007 and a PHI value of 0.06

[Zhian Kamvar]

Hi,

This question is general enough that it is outside the scope of this forum (which is mainly for procedural information on poppr and/or bugs). This question is about the output of AMOVA, which is controlled by ade4 or pegas (poppr.amova is a wrapper for both). I would encourage you to post this to the R-SIG-Genetics list: https://stat.ethz.ch/mailman/listinfo/r-sig-genetics

Regarding your question: One thing to consider is asking whether or not within individual variance is important to consider for your inquiry.

Best,

Zhian