Can't explain AMOVA result

Question

[weedy23]

Hi there, I am hoping that someone can help me explain this result I got when conducting an AMOVA analysis on SNP data in Arlequin. I understand the test and have used AMOVA before but this result has me confused. I am testing 3 populations, 2 which are grouped together. The two grouped populations show very little differentiation in other tests, while the third population is very different to the other two. This is expected as the grouped pops are very geographically close, and the third population is very distant. AMOVA gave me this:

Among groups variance (Va) = 282 (df = 1); Percentage of variation = 39.69%

Among populations within groups variance (Vb) = 1.61 (df = 1); Percentage of variation = 0.23%

Within populations variance (Vc) = 427 (df = 185); Percentage of variation = 60.1%

Fixation Indices

  FST :      0.39919
  FSC :      0.00378
  FCT :      0.39691

Significance tests (1023 permutations)

Vc and FST : P(rand. value < obs. value) = 0.00000 P(rand. value = obs. value) = 0.00000 P-value = 0.00000+-0.00000

Vb and FSC : P(rand. value > obs. value) = 0.71457 P(rand. value = obs. value) = 0.00000 P-value = 0.71457+-0.01625

Va and FCT : P(rand. value > obs. value) = 0.00000 P(rand. value = obs. value) = 0.33920 P-value = 0.33920+-0.01796

Results are as expected apart from Va and FCT being non-significant. Considering the grouping described above, and the high FCT value (0.397), I expected FCT to be significant. Can anyone explain this for me?

[pvrandsoe]

Hi there. Interesting question. I have the same problem with my current data. I do not have any clarification for you, however.

Did you find a solution or and a plausible interpretation of your AMOVA? I would realy like to hear how it turned out.

Best Phillip

[Ram]

This is more appropriate as a comment, not an answer. I'm moving it to a comment.