Whole-genome sequence analysis shows that two endemic species of North American wolf are admixtures of the coyote and gray wolf

Abstract

Protection of populations comprising admixed genomes is a challenge under the Endangered Species Act (ESA), which is regarded as the most powerful species protection legislation ever passed in the United States but lacks specific provisions for hybrids. The eastern wolf is a newly recognized wolf-like species that is highly admixed and inhabits the Great Lakes and eastern United States, a region previously thought to be included in the geographic range of only the gray wolf. The U.S. Fish and Wildlife Service has argued that the presence of the eastern wolf, rather than the gray wolf, in this area is grounds for removing ESA protection (delisting) from the gray wolf across its geographic range. In contrast, the red wolf from the southeastern United States was one of the first species protected under the ESA and was protected despite admixture with coyotes. We use whole-genome sequence data to demonstrate a lack of unique ancestry in eastern and red wolves that would not be expected if they represented long divergent North American lineages. These results suggest that arguments for delisting the gray wolf are not valid. Our findings demonstrate how a strict designation of a species under the ESA that does not consider admixture can threaten the protection of endangered entities. We argue for a more balanced approach that focuses on the ecological context of admixture and allows for evolutionary processes to potentially restore historical patterns of genetic variation.

Keywords: Admixture; canids; genomics.

Fig. 1. Admixture proportions, hypothesized branching patterns, and the geographic distribution of Canis in North America.

Top: Previously proposed phylogenetic relationships among Canis lineages, with gray lines indicating putative admixture events (5). Bottom: Geographic distributions of Canis in North America. Sample locations are indicated by dots and abbreviations are described in Table 1. Ancestry proportions from vonHoldt et al. (5) are indicated (proportion gray wolf/proportion coyote; see also new values in Table 3). IRNP, Isle Royale National Park; Ma, million years ago.

Fig. 2. PCA of 5.4 million unphased SNPs and 23 Canis genomes.

The dashed line contains genomes that are considered admixed.

Fig. 3. Estimates of ancestry proportions using the f^ statistic.

Sequences grouped as coyotes are from Alabama, California, Illinois, Ohio, and Florida. (*Individual labeled as wolf but is likely to derive from a coyote; see discussion in the text.)

Fig. 4. Demographic history inferred using G-PhoCS.

A schematic depiction of the population phylogeny assumed in the analysis. The phylogeny was augmented with migration bands from all canids to the red wolf and the Great Lakes region wolf. G-PhoCS infers significant rates of gene flow primarily from the gray wolf and the coyote to the red wolf and the Great Lakes region wolf (shaded box). Ninety-five percent Bayesian credible intervals are shown for the total rates transformed into proportions between 0 and 100% (see Materials and Methods). Similarly high rates were also inferred when assuming three alternative topologies for the population phylogeny (fig. S4).