I am guessing you mean 7 loci and not alleles. For the locus that is 100 percent homozygous, is it fixed for the same repeat in all populations? If so, that just means it is not variable and therefore doesn't provide any information.

I am guessing you mean 7 loci and not alleles. For the locus that is 100 percent homozygous, is it fixed for the same repeat in all populations? If so, that just means it is not variable and therefore doesn't provide any information.

Zero variability does not equal zero information! See, for example:

Donnelly P, Tavaré S, Balding DJ, Griffiths RC. Estimating the age of the common ancestor of men from the ZFY intron. Science 1996;272:1357-9.

or

Rodriguez R, Ramirez O, Valdiosera CE, Garcia N, Alda F, Madurell-Malapeira J, Marmi J, Doadrio I, Willerslev E, Gotherstrom A, Arsuaga JL, Thomas MG, Lalueza-Fox C, Dalen L (2011) 50,000 years of genetic uniformity in the critically endangered Iberian lynx. Mol Ecol 20:3785-3795.

Just because its a microsatellite, does not mean it can't have been influenced by selection. If a locus closely linked to this microsatellite has been under strong positive selection then that could drive the low microsatellite diversity.

It could also be that the microsatellite has a low mutation rate (for any of a number of reasons).

It could also be simple genealogical stochasticity.

Thanks for your answers.

Yes, sorry, I did mean "loci"-- I have a very bad habit of using that term interchangeably with "allele"-- even though I know it is not the same thing. Co-authors have corrected me on that misuse a bunch of times in the past.

But, I did want to clarify, the marker I'm talking about *is* polymorphic between populations-- it's just that it always seems to be homozygous. The mean Ho value is 1.00 for every population for this marker, even though mean He values are predicted as being anywhere between 0.4 and 0.6. That is really what my question concerns. I suppose, it is possible, as Mark Thomas suggested above, that there is still selection keeping this loci from altering over time.

Check wether the sequence containing the "homozygous" SSR is of nuclear or organellar origin. (If the core sequence of the marker is available you can easily decide this by BLAST-ing.) If a marker is of chloroplast or of mitochondrial origin, we expect maternal inheritance and no segregation. On the other hand, different types could be fixed in distant populations.

Colleen Ingram noted, and accurately, that if you find the same repeat in all populations, then it is not variable and provides no info. However, that is not the situation we have found. The sequence *is* variable between populations-- what is not changing is homozygosity in all five populations. So really what I am trying to figure out is what force appears to be fixing homozygosity on this one locus-- not sequence length? Because, the actual sequence length is changing between populations.

I know this may be a moot point, but zero variation is definitely not zero information in population genetics!!!

Do you mean the five populations are fixed for alternative alleles at that locus? How many alleles do you have considering all the five populations?

Ana: They range from 3-7 alleles at that locus for the five populations. There are twelve different alleles all together. Our outgroup has more alleles (7) than any of the examined populations. Again, even given all the various alleles, we have few heterozygotes at that locus.

James, if your locus is homozygote (or almost homozygote) for alternative alleles in different populations you might think that this locus has been driven by positive selection (selective sweep)

I think that one possibility is that you are dealing with allele drop out