Reflections on “Insights from population genetics for range limits of a widely-distributed annual plant”

This paper is the final of four papers (Stanton-Geddes and Anderson 2011, Stanton-Geddes et al. 2012b, 2012a) from my PhD examining range limits of the native annual legume Chamaecrista fasciculata. Upon reflection of the publication of this paper (I finished the preliminary analysis for this paper almost three years ago, and had the first draft done by my thesis defense in August 2011. Of the four papers, this one changed the most between drafts and reviews, and most fully represents my thinking on range limits in C. fasciculata), there are three aspects of this paper that I find to be generally interesting.

First, the “abundant center” model of species’ distributions has been tested in hundreds of studies using molecular markers (Eckert et al. 2008)^1^. Most of these studies used allozymes (62%) or microsatellites (23%) and few used actual sequence data. Thus, the statistic most commonly reported was expected heterozygosity (H_e). H_e is a supposedly simple statistic that’s frustratingly confusing as it depends on the number of alleles at a locus and the frequency of each allele. Thus, you can end up with low or high values depending not only on the number of alleles, but also their evenness. By using actual sequence data, I was able to calculate a much clearer statistic, nucleotide diversity which is simply the average number of nucleotide differences per site between any two DNA sequences. Interestingly, I found that while H_e did not differ among populations, nucleotide diversity was significantly higher in my “interior” population than the edge populations. So - given how cheap and easy it is to get sequence data these days, stop using microsats and collect sequence data!

Second, ecologists are often interested in whether a population is expanding or contracting. For example, if a population at a range edge is expanding, this may suggest the species is currently undergoing range expansion, but the time scale is too slow for us to observe. In contrast, a stable or contracting population is not likely expanding its range (in that direction). It’s difficult to collect enough years of demographic data for most species to assess this, which is why we turn to population genetics. In an expanding population (thus evolving non-neutrally) we will observe an excess of low frequency polymorphisms compared to expectations, which is captured by negative values of the statistic Tajima’s D. It’s worth nothing that Tajima’s D was created to identify sequences that are evolving non-neutrally, and not just due to population expansion. When I looked at Tajima’s D in the 9 sequences I had, I found no significant differences among populations or deviations from the null expectation of zero.

My PhD adviser Peter Tiffin, based on previous experience with Tajima’s D not behaving as ‘expected’ (Moeller et al. 2007), prompted me to perform simulations to evaluate what values I would expect for this statistic given what is known about the demographic history of this species. So I plugged a range of likely values into the program ms given the geologic history of the locations examined (e.g. last glaciation ~ 16k years ago) and then calculated Tajima’s D a bunch of times. From these simulations, I found that there may be no demographic signal in Tajima’s D, even if the population had experienced a substantial bottleneck followed by expansion. This occurs because the two processes, contraction and expansion, occur on different timescales and the signal from the expansion masks the contraction. While I was writing up the paper, a similar simulation result was published for the opposite scenario, showing that the demographic signal of an old expansion persists for a long time during a slow contraction (Arenas et al. 2011). However, a sobering result is that even with the size of the dataset I collected, we lacked power to detect to discard the alternatives. This is where next-gen sequencing comes in so that future studies aren’t marker limited!

Finally, at a seminar I gave to the EEB department during my PhD, paleoecologist Ed Cushing pointed out that the northern range limit of C. fasciculata in Minnesota south and west of the Minnesota and Mississippi Rivers in Minnesota is also the geographic region that would have been kept clear of trees by fires set by Native Americans. Also, Native Americans used the rivers as “highways” and often moved seed. This idea was intriguing. A little searching revealed that there were ethnobotanic uses for C. fasciculata. Further, I found that the Cahokian culture had its epicenter near modern day St. Louis (which I vaguely knew about having lived in the area), but also included settlements as far north as modern day Trempeleau, Wisconsin, where you can still find C. fasciculata today. Tantalizingly, the population structure did indicate that there may have been gene flow north from the Missouri to Iowa to Minnesota populations, but not to the west. So, the scenario I envision is that Cahokians moved the seed, which otherwise has very limited dispersal, from where it was trapped by the Last Glacial Maximum, into southern Minnesota. The plant was “pre-adapted” to environmental conditions up to Minnesota, and prospered in the areas maintained clear of forest by fire. So, while my conclusion today is that the species is in equilibrium with climate, this may only be because it was moved “everywhere” by people (before cars and planes!) and fills the ecological conditions it can tolerate. However, if I had performed my study 800 years ago, I might have found that it was not limited by climate and still had the potential to expand its range northward.

Ecologists are constantly writing about the effects that humans are (and will) have due to C0² induced climate change, forgetting that humans have been one of the major factors influencing biota since that neolithic revolution. It’s worth keeping this in mind when studying the distribution and abundance of species!

Footnotes

1: The best way to assess population size is to go out and measure it (e.g. the Christmas Bird Survey), year after year. But this isn’t possible or practical for most species. And in other cases, we’d like a longer time-scale than is possible (e.g. many generations of a long-lived species). This is where population genetics comes in. Molecular genetic diversity can be used as a proxy for effective population size. Simply put - the more diversity, the larger the population. Of course, there are all sorts of confounding factors such as history of colonization, bottlenecks, generation time and gene flow, but in general, this pattern holds up.

References

Arenas, M., N. Ray, M. Currat, and L. Excoffier. 2011. Consequences of Range Contractions and Range Shifts on Molecular Diversity. Molecular Biology and Evolution.

Eckert, C. G., K. E. Samis, and S. C. Lougheed. 2008. Genetic variation across species geographical ranges: the central–marginal hypothesis and beyond. Molecular Ecology:1170–1188.

Moeller, D. a, M. I. Tenaillon, and P. Tiffin. 2007. Population structure and its effects on patterns of nucleotide polymorphism in Teosinte (Zea mays ssp. parviglumis). Genetics 176:1799–1809.

Stanton-Geddes, J., R. G. Shaw, and P. Tiffin. 2012a. Interactions between Soil Habitat and Geographic Range Location Affect Plant Fitness. PLoS ONE 7:36015.

Stanton-Geddes, J., P. Tiffin, and R. G. Shaw. 2012b. Role of climate and competitors in limiting fitness across range edges of an annual plant. Ecology 93:1604–1613.

Stanton-Geddes, J., and C. G. Anderson. 2011. Does a facultative mutualism limit species range expansion?. Oecologia 167:149–155.