Genotyping by sequencing reveals contrasting patterns of population structure, ecologically mediated divergence, and long-distance dispersal in North American palms

Abstract

Comparative studies can provide powerful insights into processes that affect population divergence and thereby help to elucidate the mechanisms by which contemporary populations may respond to environmental change. Furthermore, approaches such as genotyping by sequencing (GBS) provide unprecedented power for resolving genetic differences among species and populations. We therefore used GBS to provide a genomewide perspective on the comparative population structure of two palm genera, Washingtonia and Brahea, on the Baja California peninsula, a region of high landscape and ecological complexity. First, we used phylogenetic analysis to address taxonomic uncertainties among five currently recognized species. We resolved three main clades, the first corresponding to W. robusta and W. filifera, the second to B. brandegeei and B. armata, and the third to B. edulis from Guadalupe Island. Focusing on the first two clades, we then delved deeper by investigating the underlying population structure. Striking differences were found, with GBS uncovering four distinct Washingtonia populations and identifying a suite of loci associated with temperature, consistent with ecologically mediated divergence. By contrast, individual mountain ranges could be resolved in Brahea and few loci were associated with environmental variables, implying a more prominent role of neutral divergence. Finally, evidence was found for long-distance dispersal events in Washingtonia but not Brahea, in line with knowledge of the dispersal mechanisms of these palms including the possibility of human-mediated dispersal. Overall, our study demonstrates the power of GBS together with a comparative approach to elucidate markedly different patterns of genomewide divergence mediated by multiple effectors.

Keywords: Arecaceae; Baja California peninsula; genotyping by sequencing (GBS); human‐mediated dispersal; single nucleotide polymorphism (SNP).

Figure 1

Maps showing the oases (points) and the sierras (shape files) from which (a) Washingtonia and (b) Brahea palms were sampled. The full names of the oases and sample sizes are given in Table S1. The colors of the shapefiles correspond to species (as defined in Minnich et al., 2011): (a) W. robusta on the Baja peninsula and Mexican mainland and W. filifera at SJ; (b) B. edulis at the Guadalupe Island (GI), B. armata at SL, SA, CAT, SSPM, and SJ and B. brandegeei at SLL, SM, SPP, and SSF

Figure 2

Bayesian reconstruction of the phylogenetic relationships (a) between Washingtonia and Brahea; (b) within Washingtonia; (c) within Brahea. Colored lines delimit morphologically defined species according to Minich et al. (2011), and colored dots correspond to the sampled populations. Major nodes with high posterior probability support (>0.95) are indicated in asterisks (*)

Figure 3

Population structure of Washingtonia palms. Panel (a) shows the results of cluster analysis within sNMF, with pie charts indicating the geographic distribution of six inferred genetic clusters (each coded by a different color). Panel (b) shows a phylogenetic network generated by SPLITSTREE

Figure 4

Population structure of Brahea palms. Panel (a) shows the results of cluster analysis within sNMF, with pie charts indicating the geographic distribution of three inferred genetic clusters (each coded by a different color). Panel (b) shows a phylogenetic network generated by SPLITSTREE

Figure 5

Maximum‐likelihood trees depicting patterns of genetic divergence among (a) Washingtonia palms grouped into eight populations and (b) Brahea palms grouped into ten populations (see Section 2 for details). Inferred migration events are indicated by dashed lines, with the direction of gene flow indicated by arrows and color intensity reflecting the intensity of gene flow

Figure 6

Phylogenetic networks constructed separately for Washingtonia using (a) neutral loci; (b) outlier loci