Tracking the origins of Yakutian horses and the genetic basis for their fast adaptation to subarctic environments

Abstract

Yakutia, Sakha Republic, in the Siberian Far East, represents one of the coldest places on Earth, with winter record temperatures dropping below -70 °C. Nevertheless, Yakutian horses survive all year round in the open air due to striking phenotypic adaptations, including compact body conformations, extremely hairy winter coats, and acute seasonal differences in metabolic activities. The evolutionary origins of Yakutian horses and the genetic basis of their adaptations remain, however, contentious. Here, we present the complete genomes of nine present-day Yakutian horses and two ancient specimens dating from the early 19th century and ∼5,200 y ago. By comparing these genomes with the genomes of two Late Pleistocene, 27 domesticated, and three wild Przewalski's horses, we find that contemporary Yakutian horses do not descend from the native horses that populated the region until the mid-Holocene, but were most likely introduced following the migration of the Yakut people a few centuries ago. Thus, they represent one of the fastest cases of adaptation to the extreme temperatures of the Arctic. We find cis-regulatory mutations to have contributed more than nonsynonymous changes to their adaptation, likely due to the comparatively limited standing variation within gene bodies at the time the population was founded. Genes involved in hair development, body size, and metabolic and hormone signaling pathways represent an essential part of the Yakutian horse adaptive genetic toolkit. Finally, we find evidence for convergent evolution with native human populations and woolly mammoths, suggesting that only a few evolutionary strategies are compatible with survival in extremely cold environments.

Keywords: adaptation; ancient genomics; horse; population discontinuity; regulatory changes.

Fig. 1.

Geographic distribution and population affinities. (A) Geographic location of ancient (blue) and modern (dark red) Yakutian horse samples (SI Appendix, Table S1.1). (B) TreeMix relationships between modern Yakutian horses (brown), ancient horses (blue), Przewalski's horses (green), and our panel of nine domestic breeds (red).

Fig. 2.

PCA. The analysis was based on genotype likelihoods and 29 genomes representative of present-day Yakutian horses, nine domestic breeds, the Przewalski’s horse population, CGG101397, and extinct horses. The fraction of the total variance explained by each of the three principal components is indicated on the corresponding axes.

Fig. 3.

Long- and short-term N_e changes of the Yakutian horse population. (A) Scaled PSMC profile tracking effective population size changes over the past 2 Mya. (B) Site frequency spectra estimated from the present-day population of Yakutian horses (including CGG101397), expected under the neutral model, and fitted under the best demographic scenario (three-epoch model). (C) Three-epoch demographic model for the Yakutian horse population. This model represents the best fit in ∂a∂i analyses. Kyr, thousand years; LGM, Last Glacial Maximum.

Fig. 4.

Relative contribution of coding and noncoding regions to adaptation. (A) Red dashed line delimits the neutrality threshold [i.e., the top-0.001% d_A quantile of 4d-fold degenerate sites (d_A = 0.3448)]. (B) Distribution of the proportion of adaptive mutations across categories, including coding (green, 0d-fold, 2d-fold, and 4d-fold degenerate positions) and noncoding (red, 10 first bins of 1 kb upstream from the TSS) sites. (C) d_A pattern at the TGM3 gene (reverse strand). (D) d_A pattern at the LECT2 and FBLX21 gene region. chr, chromosome.