Whole mitochondrial genome sequencing of domestic horses reveals incorporation of extensive wild horse diversity during domestication

Abstract

Background: DNA target enrichment by micro-array capture combined with high throughput sequencing technologies provides the possibility to obtain large amounts of sequence data (e.g. whole mitochondrial DNA genomes) from multiple individuals at relatively low costs. Previously, whole mitochondrial genome data for domestic horses (Equus caballus) were limited to only a few specimens and only short parts of the mtDNA genome (especially the hypervariable region) were investigated for larger sample sets.

Results: In this study we investigated whole mitochondrial genomes of 59 domestic horses from 44 breeds and a single Przewalski horse (Equus przewalski) using a recently described multiplex micro-array capture approach. We found 473 variable positions within the domestic horses, 292 of which are parsimony-informative, providing a well resolved phylogenetic tree. Our divergence time estimate suggests that the mitochondrial genomes of modern horse breeds shared a common ancestor around 93,000 years ago and no later than 38,000 years ago. A Bayesian skyline plot (BSP) reveals a significant population expansion beginning 6,000-8,000 years ago with an ongoing exponential growth until the present, similar to other domestic animal species. Our data further suggest that a large sample of wild horse diversity was incorporated into the domestic population; specifically, at least 46 of the mtDNA lineages observed in domestic horses (73%) already existed before the beginning of domestication about 5,000 years ago.

Conclusions: Our study provides a window into the maternal origins of extant domestic horses and confirms that modern domestic breeds present a wide sample of the mtDNA diversity found in ancestral, now extinct, wild horse populations. The data obtained allow us to detect a population expansion event coinciding with the beginning of domestication and to estimate both the minimum number of female horses incorporated into the domestic gene pool and the time depth of the domestic horse mtDNA gene pool.

Figure 1

Cladogram of the fully resolved PHYML tree estimated using the Modeltest-specified HKY+I+G substitution model, with support values from each major type of analysis displayed on each bipartition. Each bipartition shows the Bremer support, bootstrap support on the MP strict consensus tree, bootstrap support on the PHYML tree, and Bayesian posterior credibility from the MrBayes GTR+I+G analysis. A dash is shown if Bremer support is 0 or negative, if MP bootstraps are below 50%, if there was no bootstrap support for the bipartition on the PHYML tree, or if the Bayesian posterior credibility was below 0.5 for the bipartition in question.

Figure 2

95% credibility intervals on node ages as estimated by BEAST using a normally-distributed prior on the horse/ass divergence time centered at 2.25 mybp with the standard deviation set to 0.3125 my.

Figure 3

Bayesian Skyline Plot of effective population size through time based on the whole mtDNA sequence from 63 horses. The beginning of the recent effective population size expansion is marked in red (median 7,000 years BP).