I don't know if comparing Fst among species would be the best approach. Especially is you are using completely different sets of MSATs. Even if you did use the same set of MSATs in two closely related species, you would likely have major differences in the level allelic variation between species simply due to ascertainment bias. I'd suggest looking into some of the alternative standardized measures of differentiation that take differences in allelic diversity levels into account. These include Gst, Hst, Dst & Jost's D ... There is has been a lot of discussion in the literature of when to use each metric. This book chapter by Chris Bird is a good discussion of the topic with a pretty comprehensive list of references.

http://www.epscor.hawaii.edu/system/files/er_book/2012/05/Birdetal_Detecting_Measuring.pdf

AS FAR AS I UNDERSTAND ....genetic diversity at the level of nucleotide variation/ similarity can be best explained using these software; i.e. MEGA, DNAsp and PAUP. All these software has lots of inbuilt parameters to analyze the genetic diversity across the species/strain/genera. and most importantly MEGA and DNASP are very user friendly.

I have no direct answer to this question, but I have an idea of a possible approach. I would try to see the partition of diversity within and among (sub)-populations and connect ot with specific characteristics of the species. See Finkeldey 1994 (and some following papers of the same author) on partition of diversity and Nybom 2004 on the diversity patterns of different markers in plant populations.

http://link.springer.com/article/10.1007%2FBF00225141?LI=true

http://www.ecologia.unam.mx/laboratorios/evolucionmolecular/images/file/ClaseBiolEvolPlantas/nybom.pdf

A qucik answer : can you precise ? what is your aim to compare population differentiation among species ?

Fst is informative within species and between populations ("pop A and pop B are more differentiated than pop A and C").

Between species I am not sure that you can compare Fst or any other genetic distance.

But you can compare the correlation between a factor and the genetic distance (pairwise Fst or anyelse). For example, two species in a same landscape : is isolation by distance higher or lower in species A by comparison to species B ?

Hi Joshua:

Your question reminds a paper that I read some time ago, It refers to the problem to asses the genetic diversity of a single species and determine if the values obtained is higher or lower. To resolve that problem, they compare the genetic diversity with other values already published from other birds. Due the size of the sequences was not the same they did a correction.

I hope this helps you

Lawrence, Hayley A., Graeme A. Taylor, Craig D. Millar, y David M. Lambert. «High Mitochondrial and Nuclear Genetic Diversity in One of the World’s Most Endangered Seabirds, the Chatham Island Taiko (Pterodroma Magentae)». Conservation Genetics 9, n.o 5 (1 de octubre de 2008): 1293-1301. doi:10.1007/s10592-007-9471-3.

I don't know if comparing Fst among species would be the best approach. Especially is you are using completely different sets of MSATs. Even if you did use the same set of MSATs in two closely related species, you would likely have major differences in the level allelic variation between species simply due to ascertainment bias. I'd suggest looking into some of the alternative standardized measures of differentiation that take differences in allelic diversity levels into account. These include Gst, Hst, Dst & Jost's D ... There is has been a lot of discussion in the literature of when to use each metric. This book chapter by Chris Bird is a good discussion of the topic with a pretty comprehensive list of references.

http://www.epscor.hawaii.edu/system/files/er_book/2012/05/Birdetal_Detecting_Measuring.pdf

A generalized measure of genetic diversity within a (sub-) population is [pi]. It is defined as the average chance to observe a difference between two allelic sites. It typically applies do single-nucleotide differences.

The problem is that variation of microsattellites is much different from average site variation (i.e. much higher). Also, microsattellite length affects its variation - hardly in a predictable manner. So, I don't see a solution to transfer microsattellite variance to standardized measure. You probably ought to sample variation in a more representative manner. RAD tag sequencing would be a technique to assess variation of an almost random subset of the genome.

As what others already mentioned, it is almost impossible to compare microsatellite datasets of different species given the differences in how the markers were screened and developed (e.g. null allele or drop-out for across species amplification), marker list (level of polymorphism of each marker affects the number of alleles observed), population structure and sampling strategy etc... We even failed to combine microsatellite datasets of the same species but generated in different labs for different populations/breeds. It only makes sense to compare the diversity estimates, e.g. the number of alleles for microsatellite markers (e.g. heterozygosity and PIC are not sensitive to the change in allele numbers) and F-statistics across populations/breeds within a species when the same set of markers are applied. No standard indicator is available for this type of markers.

Hi Joshua,

My understanding is that you can compare genetic diversity between different populations, mostly when you are using same molecular marker (i.e. SSR). I've worked on some Cactaceae species, evaluating their genetic diversity. Using same SSR in different populations of same species it's possible to find a similar scenario like when you compare between species (something that you might find if you use a different set of SSR too). It's recommended to analyse your data per population, then calculate the species genetic diversity using an unbiased estimator for gene diversity (which allows you to compare them between species).

I can tell that if you use different kind of molecular markers, their biological nature would make a value comparison between species less accurate. However, you will find that in many cases, those are the only available values therefore, the only reference to compare with. I'm sure that any further insight will be very useful.

You may want to consult this paper: Hedrick, P. W. (2005). A standardized genetic differentiation measure. Evolution, 59(8), 1633-1638.

I wonder why did you ask this question. Maybe you detected loci with different degree of genetic variability in different species? Anyway, if you have a sufficient number of loci for each species, then you can compare estimators of genetic diversity directly, because you assume that loci are representative of the entire genome. Instead, I believe that comparisons among estimators obtained with different markers do not make sense.

Thank you all for your contributions! The feedback has been really good.

Just to clarify in case some people didn't quite get why I asked the question: I understand that different molecular markers - and indeed different loci of the same molecular marker type - have different evolutionary histories, and this means that a direct comparison of these markers is not necessarily valid because of this. Thus, I wanted to know if anyone knew of a way to directly compare genetic diversity in different species that have been measured with different molecular markers.

Again, thanks! I'll make an effort to follow the citation trail that people have left for me; you are all helping me learn and grow as a budding young scientist!

- Josh

The issue is about sampling the genomes - you want to make sure the genomes of both species are sampled in a comparable way. Would be possible by using high numbers of microsatellites for instance (selected without the mentioned ascertainment bias). And although the use of "random markers" should really be discouraged strongly, I wonder whether, in this case only, a high number of those wouldn't also give you some clue.

hi Joshua,

you should use variable sites (markers) that were polymorphic in the ancestor before species spit and that are still variable in the daughter species. Microsatellites are hardly in this situation. Moreover there is a huge ascertainment bias as they are often discovered in a single species and then adapted to others. SNPs could be better suited, provided the conditions above. cheers alessio