2014-05-05_REVISION - C1996083 - Laserfiche WebLink

Home Browse Search

HISTORICAL DNA REVEALS EXTINCTIONS AND INVASIONS 7 (A) (B — N .. 0 Yampa River South Platte River -�_ O C} p Colorado River 0 Gunnison River � Arkansas River 411 —• Rio Grande River # fish o O r Gan lean River Y _ 1 • Fig. 4 (A) Statistical parsimony haplotype network using a subset of 430 basepairs of combined ND2 and COI mitochondrial sequence data. Haplotype colours match those shown in Fig. 2. Haplotypes present in (i) museum samples only have a black outline (ii) museum and modern samples have a grey outline and (iii) modern populations only have no outline. Putative unsampled ances- tors are represented by small, open circles. Numbers next to museum haplotypes refer to fish sample numbers in Table St. (B) Geo- graphical distribution of museum samples, with the colour of each circle corresponding to the lineage colour in the haplotype network. The size of each circle is proportional to the number of museum individuals sampled with a particular lineage. Colours for drainages correspond with our hypothesis of the historical distribution of clades. least two and as many as 12 mutations ( -0.5 to 3 % sequence difference) (Fig. 4A). Six of the 15 haplotypes represented in the museum samples were identical to modern haplotypes, and nine were unique (Fig. 4A). Importantly, museum samples from the Arkansas and San Juan basins did not group with any of the contem- porary lineages, and the two sets of haplotypes were separated by a minimum of four mutations from any other clade (Fig. 4A). Our phylogenetic inference using maximum- likeli- hood and Bayesian analyses included samples represen- tative of contemporary and historical subspecies in Colorado as well as cutthroat trout subspecies and rain- bow trout native to drainages outside of the state. The combination of modern and museum samples revealed high maximum - likelihood ratios and posterior probabil- ities (using both tree priors —a coalescent model assum- ing constant population size and a Yule speciation process model) for eleven North American cutthroat trout clades. Of the six clades represented by fish col- lected from Colorado, two were only detected in museum samples from either the San Juan or Arkansas drainages (Fig. 5). Furthermore, twenty -one museum samples grouped within four contemporary clades with high posterior probabilities in a phylogenetic tree (Fig. 5). Importantly, none of the branches subtending two or more clades had consistently high support across phylogenetic analyses, underscoring that the rela- tionships among clades remain uncertain. If branching © 2012 Blackwell Publishing Ltd order disagreed between tree inference methods, a polytomy was enforced (Fig. 5). Bayesian posterior probabilities and approximate likelihood test ratios were reported for all other branches. For the Bayesian analyses, posterior probabilities were highly similar ( <1% in most cases) regardless of tree prior, and there- fore, only the values from the coalescent analysis are shown in Fig. 5. Phylogenetic analyses reveal historical distribution of Colorado's cutthroat trout There were six divergent mtDNA clades discovered from sequencing of the museum samples —of these, four were restricted to single drainage basins (Fig. 413). All historical samples from the Rio Grande River drain- age grouped with the lineage designated Rio Grande cutthroat trout (orange lineage). All samples that grouped with the purple lineage were restricted to the South Platte drainage. The two distinct clades that did not cluster with any of the modern clades (shown in red and yellow) were restricted to either the San Juan basin or the Arkansas basin, respectively (Fig. 413). The blue and green lineages, however, were discovered in multiple drainages. It is noteworthy that the museum samples comprising these two lineages were sampled in 1889, after the initiation of stocking activities in the state. Therefore, their native status is somewhat uncer- tain. The two museum samples that grouped with