Hi Jana,

AMOVA does not calculate difference among groups if you have only 2 populations. 

You need, at least, one group with two different populations. Example: (Pop1+Pop2) x Pop3.

If you want to suggest Management Units (short term conservation actions) in your research you should not use information about distance, only haplotype frequencies. If you want to incorporate information on evolutionary levels (ESU), you should use distance! Have a look at this: ˜Moritz, C. 1994. Defining ‘Evolutionary Significant Units’ for consrvation. Trends in Ecology & Evolution. v.9. 10: 373-375.˜

Hope it helps!

Cheers,

Sarah

The Analysis of Molecular Variance (AMOVA) procedure follows the methods of Excoffier et al. [1992], Huff et al. [1993], Peakall et al. [1995], and Michalakis and Excoffier [1996]. AMOVA allows the hierarchical partitioning of genetic variation among populations and regions and the estimation of the widely used F-statistics and/OR THEIR ANALOGUES. The AMOVA framework was a key development in population genetic analysis, because it allowed (for the first time) hierarchical population structure analysis for all types of genetic markers (HAPLOID, HAPLOTYPE, binary, codominant and sequence), as well as offering statistical testing options by random permutation. There have been a number of later variations on this same theme.

Within GenAlEx (FREE SOFTWARE), the data type and choice of distance calculation used as input for AMOVA lead to related but different analyses. To estimate FST for codominant data, choose the Codom-Allelic distance. To estimate RST for microsatellites, choose Codom-Microsat. Note, RST should not be your default option for microsatellites, because the underlying assumptions of simple step-wise mutation rarely hold in natural populations. Consequently, estimation of RST is often not useful. To suppress within population variance and simply calculate population differentiation based on the genotypic variance, choose Codom-Genotypic distance. This option produces an estimate of ΦPT, an analogue of FST. ΦPT is also the estimate of population genetic differentiation provided by GenAlEx when binary or HAPLOID data are analysed. When comparing patterns of molecular variance between codominant and other markers (such as binar  AFLPs), ΦPT should be used for all markers (see [Peakall et al., 1995] and [Maguire et al., 2002] for comparative studies) In all other cases involving codominant data, we recommend reporting FST rather than ΦPT. GenAlEx 6.5 offers new AMOVA routines to estimate standardized FST, following Meirmans [2006], and we suggest this statistic should also be reported in all studies. The software packages Arlequin [Excoffier and Lischer, 2010] and Genodive [Meirmans and Van Tienderen, 2004] also offer AMOVA options. GenAlEx offers data export to both packages. See also, F-statistics, G-statistics and Genetic Distance below.

I agree with the previous answer, preferably you need more than two populations. But with GenAlEx can explore a first approximation. Do not use FST, for this type of data is better ΦPT.

Cheers, 

Pablo

Hi Jana,

Yes you need more population and the frindly programme to use is GenAlEx as Pablo said see the link below.

http://biology-assets.anu.edu.au/GenAlEx/Welcome.html

Best wishes,

Ali

Hi all!

Thank you for your help. I would like to clarify my question. I tested 4 distinct colonies equally split between 2 regions. I erroneously referred to each region as a 'population'. Nevertheless, your support has helped me a lot.

Cheers,

Jana