Since you’re reading this blog, you must know that I don’t have a lot of time for research these days. My duties as Vice Provost for Graduate Education and Dean of The Graduate School at UConn take up most of my time. I do manage to contribute to some research, so long as other people do the real work and I contribute some ideas or some statistical analyses. Here’s another example of that.
Last fall I was asked about the old C++ program Hickory that I had written to facilitate analysis of Wright’s F-statistics with dominant markers. It was never terribly widely used, and it was difficult to maintain. I gave up about 10 years ago. In the meantime, I realized that there’s an easy way to rewrite Hickory using Stan. After being contacted, I finally bit the bullet and did the rewrite in a combination of Stan and R. I even mentioned the R/Stan implementation last September.
Yesterday, we posted a pre-print on bioRxiv that uses the new version of Hickory as one of a variety of analytical methods that provide some insight into the genetic structure of Leymus chinensis. Here’s the abstract and a link.
Genetic structure in patchy populations of a candidate foundation plant: a case study of Leymus chinensis (Poaceae) using genetic and clonal diversity
Jian Guo, Christina L. Richards, Kent E. Holsinger, Gordon A. Fox, Zhuo Zhang, Chan Zhou
PREMISE The distribution of genetic diversity on the landscape has critical ecological and evolutionary implications. This may be especially the case on a local scale for foundation plant species since they create and define ecological communities, contributing disproportionately to ecosystem function.
METHODS We examined the distribution of genetic diversity and clones, which we defined first as unique multi-locus genotypes (MLG), and then by grouping similar MLGs into multi-locus lineages (MLL). We used 186 markers from inter-simple sequence repeats (ISSR) across 358 ramets from 13 patches of the foundation grass Leymus chinensis. We examined the relationship between genetic and clonal diversities, their variation with patch-size, and the effect of the number of markers used to evaluate genetic diversity and structure in this species.
RESULTS Every ramet had a unique MLG. Almost all patches consisted of individuals belonging to a single MLL. We confirmed this with a clustering algorithm to group related genotypes. The predominance of a single lineage within each patch could be the result of the accumulation of somatic mutations, limited dispersal, some sexual reproduction with partners mainly restricted to the same patch, or a combination of all three.
CONCLUSIONS We found strong genetic structure among patches of L. chinensis. Consistent with previous work on the species, the clustering of similar genotypes within patches suggests that clonal reproduction combined with somatic mutation, limited dispersal, and some degree of sexual reproduction among neighbors causes individuals within a patch to be more closely related than among patches.