Chromosome-scale Genomes Show Rapid Diversification and Ancient Gene Flow Among Bear Species

Abstract

Reconstructions of evolutionary history can be restricted by a lack of high-quality reference genomes. To date, only four of the eight species of bears (family Ursidae) have chromosome-level genome assemblies. Here, we present assemblies for three additional species-the sun, sloth, and Andean bears-and use a whole-genome alignment of all bear species and other carnivores to reconstruct the evolution of Ursidae. Multiple divergence dating approaches suggest that the six Ursine bears likely diversified in the last 5 Ma, but that divergence times within Ursinae are significantly impacted by gene tree heterogeneity. Consistent with this, we observe that nearly 50% of gene trees conflict with our highly supported species tree, a pattern driven by a significant early hybridization event within Ursinae. We also find that the karyotype of Ursinae is largely similar to the ancestral karyotype of all bears twenty million years prior. In contrast to this conservation of structure, dozens of chromosomal fissions and fusions associated with LINE/L1 retrotransposons dramatically restructured the genomes of the giant panda and Andean bear. Finally, we leverage these genomes to identify species-specific evidence for positive selection on genes associated with color, diet, and metabolism. One of these genes, TCPN2, has a role in pigmentation and shows a series of amino acid mutations in the polar bear over the last 0.5 Ma. Collectively, these new genomic resources enable improved reconstruction of the complex evolutionary history of bears and clarify how this enigmatic group diversified.

Keywords: bears; chromosome evolution; gene flow; phylogenomics; positive selection.

Fig. 1.

Coalescent-based model of divergence times for Ursidae. Gray bars at nodes indicate 95% posterior distributions. Map inset displays the current ranges of each species. Sillhouttes from PhyloPic.

Fig. 2.

a) Site and gene concordance factors calculated from maximum-likelihood trees based on 50 kb windows. Each point represents the concordance factor value for a single chromosome associated with a given node in the species tree. b) Genome-wide frequencies of the heaviest-weighted topology of three alternatives (TWISST) for the same set of 50-kb windows. c) Estimated likelihood for networks with a range of 0 to 8 hybrid edges, where lower likelihood indicates a better fit. d) Best-fit explicit phylogenetic network for 1 hybrid edge as inferred from 50 kb windows.

Fig. 3.

Chromosome rearrangements during the evolution of bears. Each node number corresponds to an ancestor in the phylogeny and its reconstructed ancestral genome, as inferred by DESCHRAMBLER (Kim et al. 2017). Reconstructed chromosomes are ordered and colored by synteny with the earliest common ancestor (Node 1). Synteny in the X chromosome is highlighted with a darker gray band.

Fig. 4.

Results of aBSREL analysis of positive selection. a) Species tree annotated with the total number of genes with evidence of positive selection on each branch. Branch color is scaled by the number of genes under selection. b) Summary of candidate genes under positive selection for each branch. Dots indicate the number of genes belonging to a given gene ontology (GO) term. Dots are colored by a broader functional grouping of candidate gene GO terms.

Fig. 5.

Structure and alignment of the TCPN2 gene in bears. a) Simple schematic of the TCPN2 protein showing the two ion transport domains. b) Alignments highlighting all positions where the polar bear differs from the consensus amino acid sequence. Dots indicate a shared amino acid with the consensus sequence.