Computational protocol: Molecular genetic analysis of two native desert palm genera, Washingtonia and Brahea, from the Baja California Peninsula and Guadalupe Island

Similar protocols

Protocol publication

[…] DNA extraction was performed using QIAGEN DNeasy plant mini kit (Valencia, CA, USA). As we are unaware of any population‐level genetic studies of Washingtonia or Brahea, we first evaluated three nuclear and seven chloroplast loci that have previously been used in phylogenetic and biogeographic studies of Arecaceae (Bacon, Baker, & Simmons, ; Bacon, Feltus, Paterson, & Bailey, ; CBOL Plant Working Group ; Jeanson, Labat, & Little, ; Shaw et al., ; Taberlet, Gielly, Pautou, & Bouvet, ). These loci (ITS 2, CISP 4, CISP 8, matK, rbcl, trnT‐trnD, trnG‐trnS, trnC‐rpoB, trnF‐trnL, trnfM‐trnS) were selected based on the amount of intrageneric diversity they possess in other plant species. For this trial, 40 samples of five palm species were PCR amplified in a 12.5 μl reactions containing 50–100 ng of DNA template, 1× PCR buffer, 1.5 mM MgCl2, 200 mM of dNTP mix, 0.4 mM of each primer, and 0.5 U of GoTaq Flexi DNA polymerase (Promega). The loci were amplified using primers and thermocycling protocols described elsewhere (Bacon et al., ; Jeanson et al., ; Shaw et al., ; Taberlet et al., ). Five loci (one nuclear and four chloroplast) were deemed suitable for further study and these were sequenced in all of the samples using the PCR conditions described above. The presence of amplified target DNA fragments was verified on 1.5% agarose gels stained with ethidium bromide. All positive PCR products were sent to Arizona GeneCore for sequencing in both directions. The resulting sequence chromatograms were assembled and manually edited using BioEdit 7.0.0 (Hall, ). The chloroplast sequences were then concatenated into a single sequence, as were the nuclear sequences, allowing subsequent analyses to be carried out separately for these two types of loci. [...] Our sample size was somewhat restricted due to the difficulty of collecting samples from these locally rare and geographically restricted endemic species. We therefore tested if our final sample size was sufficient to recover most of the genetic diversity within the sampled populations. For this analysis, we constructed haplotype accumulation curves separately for the nuclear and chloroplast sequence data using the function haploAccum in the R package SPIDER (R Development Core Team 2008) with a total of 10,000 permutations. Molecular diversity statistics including the number of variable sites, the number of haplotypes, nucleotide and haplotype diversities, and the number of shared haplotypes were then computed using DnaSP 5.1 (Librado & Rozas, ) and ARLEQUIN 3.5.1.3 (Excoffier & Lischer, ). Phylogenetic relationships among the samples were inferred on the basis of the nuclear and chloroplast sequence data by constructing a parsimony network (Clement, Posada, & Crandall, ; Templeton, Crandall, & Sing, ) in PopArt (http://popart.otago.ac.nz, June 2016). We also used Gengis 2.1.1 (Parks et al., ) to visualize plastid and nuclear sequence geographic distributions.To assess levels of genetic divergence between putative taxonomic species and a priori defined geographical regions (Figure , Table  and see below), we quantified pairwise differentiation using F st (Weir & Cockerham, ) and Φst (Kimura, ). The former looks only at haplotype frequency differences (Excoffier, Smouse, & Quattro, ) while the latter also incorporates haplotype sequence similarity. F st and Φst values were calculated within ARLEQUIN. For Washingtonia, eight geographic regions, each corresponding to a different sierra, were defined as follows: Mexican mainland (Son), Cabo Region (inluding Sierra La Laguna combined with Sierra Cacachilas, SL), Sierra del Mechudo (SM), Sierra la Giganta (SG), Sierra San Pedro (SSP), combined sites of Sierra Libertad and Sierra San Francisco (SLI + SSF), Cataviña (Cat) and Sierra Juárez (SJ). For Brahea, we specified ten geographical regions. Nine of these corresponded to different sierras on the Baja Peninsula: Cabo Region (SL), Sierra del Mechudo (SM), Sierra San Francisco (SSF), Sierra San Pedro (SSP), Sierra Libertad (SLI), Sierra Asamblea (SA), Cataviña (Cat), Sierra San Pedro Martir (SSPM) and Sierra Juárez (SJ); while the final geographical region corresponded to Guadalupe Island (GI). For further details including sampling coordinates, see Tables  and ; Figures  and .We also used hierarchical analysis of molecular variance (AMOVA) to partition the genetic variation within Washingtonia and Brahea using ARLEQUIN. We used AMOVA to estimate relative support for three alternative scenarios (Figure ). Each scenario specified a hierarchical partitioning of populations within regions/groups, and the nesting scheme that maximized the among‐group variance component was considered the best divergence scenario. The first scenario was based on the current taxonomical delimitation of the species (i.e. two groups were specified for Washingtonia, corresponding to W. filifera and W. robusta and three groups were specified for Brahea corresponding to B. edulis, B. armata, and B. brandegeei). Second, for the Washingtonia palms, we constructed a scenario based on the isolation of the mainland and peninsular populations after the formation of Sea of Cortez into three major regions: the Mexican mainland, Baja Peninsula and samples of W. filifera. Third, for the Brahea palms, we partitioned all of the samples into three groups: Guadalupe Island, Cabo region and rest of the Baja Peninsula. Fourth, for both Washingtonia and Brahea palms, we tested the hypothesis that relict populations persisted within each of the sierras, which should be reflected in each sierra harboring a genetically distinctive population. Eight groups were specified for Washingtonia and ten for Brahea, each corresponding to a different sierra (For more details see Figure ). AMOVA and genetic differentiation analyses were not carried out for the Brahea chloroplast data, as only three haplotypes were present and almost all of the individuals shared the same haplotype. The single B. elegans sample was also excluded from these analyses. For each AMOVA, statistical significance was determined using 10,000 permutations of the dataset.Finally, we used a Bayesian approach to estimate the number of genetic clusters present within the data as implemented in BAPS 6 (Corander, Marttinen, Siren, & Tang, ; Corander, Siren, & Arjas, ; Corander, Waldmann, & Sillanpaa, ). For this analysis, individuals were first clustered without any previous knowledge of their sampling locations. This analysis was run using the mixture model to determine the most probable number of clusters (K) within the data. K was set from 1 to 10 and five replicates were performed for each value of K. An admixture analysis was then carried out with 50 reference individuals and 1,000 iterations. Finally, we spatially clustered the individuals using Voronoi tessellation and geo‐referenced sample location data. All of the spatial analyses were conducted separately for Washingtonia and Brahea. […]

Pipeline specifications

Software tools BioEdit, DnaSP, Arlequin, PopART, GenGIS, BAPS
Applications Phylogenetics, Population genetic analysis
Diseases Adenoma, Islet Cell