Генетики
из венского университета Ветеринарной Медицины выяснили, что почти все
современные породы лошадей имеют общих предков, завезенных в Европу с Востока около семисот лет назад. Две основные клады
предков представлены восточной арабской линией и туркменской линию. Знаменитые
английские чистокровные верховые лошади оказались потомками туркменских лошадей.
Генетика лошадей
представляет в современном мире большой интерес не только для исследователей, но и для заводчиков. Люди, занимающиеся
разведением лошадей, например, для скачек, хотят знать, какие скрещивания
окажутся наиболее перспективными — по результатам генетического анализа можно предсказать такие характеристики,
как окрас, размеры, выносливость, сообразительность и характер будущих жеребят.
Происхождение пород лошадей также можно узнать с помощью исследования ДНК, и здесь на руку играет тот факт, что в течение сотен лет многие заводчики вели
детальные родословные своих животных — ни о каких других породах и видах животных таких подробных записей в мире, наверное, не существует.
Низкая
вариабельность Y-хромосомы у коней затрудняет проведение генеалогического анализа, однако в данном
проекте ученые воспользовались данными по длинным участкам Y-хромосом с повторами, принадлежащих 52 коням из 21 породы, и их оказалось достаточно для построения филогенетических деревьев. Деревья,
в свою очередь, сверяли с генеалогическими записями последних нескольких веков. Впоследствии анализ
некоторых участков был расширен до 363 коней из 57 пород. В качестве аутгрупп (контроля) использовались лошадь
Пржевальского и осел.
Выяснилось, что все
современные породы, по-видимому, имеют общих предков, около семисот лет назад
завезенных в Европу с Востока. В образовании новых пород участвовали две
основные клады — арабские лошади и туркменские лошади.
Шотландский пони, норвежская фьордовая лошадь и исландская лошадь оказались наиболее далекими родственниками остальных
лошадей. Лошадей привозили, например, в качестве подарков правителям, кроме
того, они служили предметом торговли. Арабские и туркменские лошади, по-видимому, давали наиболее плодовитое потомство,
поэтому их раз за разом скрещивали с другими лошадьми, в результате чего они стали основной составляющей мужской линии современных
пород. Среди потомков туркменских лошадей — в частности, английская чистокровная
верховая, южно-германская тяжелоупряжная, американская Quarter Horse и Аппалуза.
Анна Казнадзей
Y Chromosome Uncovers the Recent Oriental Origins of Modern
Stallions
Highlights
- · Y chromosomes of modern horse breeds arose from a single
ancestor after domestication
- · Sex-biased selection increased a few Oriental-derived Y chromosome lineages
- · English Thoroughbred founder stallions can be traced back to a Turkoman origin
Summary
The Y chromosome directly reflects male genealogies, but the extremely
low Y chromosome sequence diversity in horses has prevented the reconstruction
of stallion genealogies. Here, we resolve the first Y chromosome genealogy of
modern horses by screening 1.46 Mb of the male-specific region of the Y
chromosome (MSY) in 52 horses from 21 breeds. Based on highly accurate pedigree
data, we estimated the de novo mutation rate of the horse MSY and showed that
various modern horse Y chromosome lineages split much later than the
domestication of the species. Apart from few private northern European
haplotypes, all modern horse breeds clustered together in a roughly
700-year-old haplogroup that was transmitted to Europe by the import of
Oriental stallions. The Oriental horse group consisted of two major subclades:
the Original Arabian lineage and the Turkoman horse lineage. We show that the
English Thoroughbred MSY was derived from the Turkoman lineage and that English
Thoroughbred sires are largely responsible for the predominance of this
haplotype in modern horses.
Results and Discussion
The male-specific region of the Y chromosome (MSY) in mammals is
transmitted without recombination from fathers to sons and reflects the
migratory and demographic history of males exclusively. Using mtDNA as the
female counterpart, it is possible to contrast the demographic history of males
and females of a single species. The horse (Equus caballus) provides a
particularly striking example of different demographic patterns between the two
sexes. In contrast to the high mtDNA diversity, with coalescence times clearly
pre-dating domestication, the MSY has extremely low diversity. The low MSY
diversity cannot be explained by a low mutation rate; the Y chromosome lineages
of modern horses are clearly distinct from those of the Przewalski’s horse (E. przewalskii),
and prehistoric horses have much more diversity. Rather, the presence of only
six Y chromosome haplotypes (HTs) in modern European horse breeds and the
limited microsatellite variability suggest an extremely low effective
population size of males. The decline of Y chromosome diversity in horses
likely started about 5,500 years ago with genetic bottlenecks during the
domestication process and was further enhanced by multiple prehistoric and
historic waves of migration. Most so-called “modern horse breeds” are the
result of centralized and organized horse breeding over the past few hundred
years. During this period, inbreeding and line-breeding concepts became
popular, and the entire horse population has been strongly affected by these
strategies.
Of particular importance was the trend to import stallions from foreign
studs to improve local herds. In central Europe, this practice started in the 16th century with the popularity of
Spanish and Neapolitan stallions. Until the end of the 18th century,
the Central European horse population was shaped by the introduction of
“Oriental stallions,” and during this period imports were largely restricted to
Turkoman (from the steppes of central Asia) and Arabian (from the Arabian
Peninsula) stallions. Stallion-mediated improvement peaked in the 19thand 20th centuries
with the enormous influence of the English Thoroughbred. The English
Thoroughbred has had a closed studbook since 1793 and was founded by an earlier
introgression of Oriental stallions, bred to local mares. This breeding history
has led to a situation in which only a handful of founder lineages remain
within modern horse breeds and the breeding success of imported bloodlines
might have resulted in the complete replacement of autochthonous Y chromosome
variants.
Recent founder effects have therefore had a major impact on MSY
diversity in horses. Using high-resolution MSY haplotyping, we aimed to unravel
the origin of influential founder stallions and to determine their genetic
influence on contemporary horse populations.
Human studies have shown the potential to screen large regions of the
MSY in multiple samples to detect single nucleotide variants (SNVs) and small
insertion/deletions (indels) and thus to derive detailed MSY genealogies. Its
highly repetitive structure makes the Y chromosome the most challenging
mammalian chromosome to sequence and assemble, and the MSY sequence is nearly
complete for only a few species. Nevertheless, Y-linked regions can be partially
assembled using next-generation sequencing data even if they are surrounded by
repetitive DNA.
We generated a horse MSY reference sequence and subsequently
used it for variant calling using short-read data from multiple individuals. To
generate the reference, we initially enriched for male-specific reads by
mapping paired-end Illumina reads from the genomes of three male Lipizzan
stallions to the published female horse reference. We then performed a de novo
assembly of all unmapped reads (∼0.7%
of the total reads) and obtained contigs with a total length of 13.6 Mb (Figure S1A). Because the reference assembly is based on an English
Thoroughbred, the de novo contigs are a mosaic of MSY sequences and
Lipizzan-specific autosomal and X chromosome insertions. Very stringent
filtering criteria were applied to extract male-specific sequences. We mapped
Illumina reads from 27 male horses of different breeds and from five females to
the de novo assembled contigs. A contig was defined as MSY-linked if it was
covered by male reads and not by female reads. We obtained 2,794 preliminary
MSY contigs covering a length of 1.67 Mb. Details of mapping coverage are given
in Figure S1B and Data S1. We scrutinized our pipeline by validating 84 of the 2,794 Y
contigs by PCR using male and female horse DNA as templates and found that 100%
of our MSY contigs were male-specific (Figures S1D–S1F). In the final filtering step, multi-copy contigs were
removed. Based on a mean normalized average coverage of ≤1.5 (Figure S1C), we obtained a reference of 2,491 high-quality single-copy
(scpMSY) contigs covering 1.46 Mb (see workflow including main results in Figure 1).
To detect variants, we mapped
whole-genome next-generation sequencing data for 52 male domestic horses from
21 different breeds to the nonrepMSY reference, including a Przewalski’s horse
and a donkey sample as outgroups. Using haploid variant calling, we considered
all SNVs and small indels on the scpMSY with at least 2-fold coverage. We
called 867 variants in domestic and Przewalski’s horse samples. In total, 53
domestic (50 SNVs and three indels) and 284 Przewalski’s horse (271 SNVs and 13
indels) variants passed several filtering steps and were categorized as “true”
variants (Data S2). Fifty-two domestic and 12 out
of 12 randomly chosen Przewalski’s horse variants were confirmed by independent
validation (Figure S2B). The coverage for the donkey
sample was fragmented, with only 80% of the scpMSY region covered. We therefore
only considered the donkey to determine the ancestral state of horse SNVs and
small indels. Together with four published MSY variants, the 49 SNVs and three
indels formed 24 individual HTs in our sample of 52 modern horses (Data S3).
A maximum-parsimony tree was built based
on inferred domestic horse HTs using the Przewalski’s horse and the donkey
sequence as outgroups (Figure 2). The two deepest splits in the
MSY ancestry separated northern European samples with two well-supported
branches: N (Shetland pony and Norwegian Fjord horse) and I (Icelandic horse).
Two sequence variants (rAX and rAY) defined a crown group, which was a sister
group to I and contained 47 samples. Within the crown group, we observed a
polytomy with four branches (A, L, S, and T). The Arabian horse, two
Arabian-influenced Trakehner stallions, a South German draft horse, and Connemara
ponies were included on branch A. Stallions with an Iberian origin, namely
a Lipizzan sire lineage with documented Spanish ancestry and a Sorraia male,
were located on branches L and S. Branch T was defined by the SNV rA and
corresponded to the previously described HT2. This group contained more than
two-thirds of our samples (37), all of which have a documented English
Thoroughbred paternal ancestry, except the Franches-Montagnes horse (Tu).
Nucleotide diversity
was extremely low (Watterson’s θ: 7.9 × 10−6 for
all domestics, 4.8 × 10−6 for the crown group), suggesting
a recent common ancestor for all HTs. With the caveat of the possible
underestimation of branch lengths due to missing variants, our whole-genome
sequencing approach provides a SNV set that can be used to date the nodes in
the genealogy without ascertainment bias. We estimated the mutation rate of the
horse MSY using four variants (rO, rE, rF, and rL) that occur in documented
pedigrees (Figures S3A–S3C). Based on
1.36 Mb MSY sequence screened and 101 generations covered by the pedigree, we
inferred a mutation rate of 2.91 × 10−8/bp/generation, an
estimate that agrees well with the rate observed in the human MSY. Assuming a
mean generation time of 7 years, we dated the most recent common ancestor
(TMRCA) of the crown group (A-L-S-T) to 647 ± 229 years ago. The entire
tree coalesced as recently as 1,328 ± 380 years ago, and TMRCA of
Przewalski’s horse and domestic horse Y chromosomes was dated to 23,716 ±
1,975 years before present (Figure 3A). Because our
estimated mutation rate per generation was based on only four mutations, it is
not very robust. Furthermore, estimates of the generation times of horses are
quite variable, resulting in a wide range of TMRCA estimates (Figure 3B). Within the
limits of our assumptions, all estimates were consistent with the view that the
modern horse MSY lineages arose from a single founder that lived after the
domestication of the species.
To address whether the observed phylogeny
covers the male breeding stock of modern breeds, we evaluated the MSY HT
distribution by genotyping 56 MSY variants in a comprehensive set of 363
males representing 57 modern breeds. A pedigree-based sampling strategy
was used to cover the influential lines of a given breed and to avoid
oversampling of relatives. A detailed list of samples is given in Data S4. We detected 36 HTs; Figure 4A gives the HT network rooted with the Przewalski’s horse.
Even in the larger sample, haplogroups N and I were restricted to northern
European breeds, with the single HTs Nf in the Norwegian Fjord horse, Ns-1
in a Swedish Coldblood horse, and Ns-2/3 in Shetland ponies. Particularly
remarkable is the high HT diversity in the Icelandic horse (I-1 to I-4).
All remaining horse breeds clustered
within the A-L-S-T crown group and mapped either to recent branches or to the
basal node (X). This implies that MSY diversity in modern horses was not
underestimated due to ascertainment bias for MSY variants. Grouping breeds by
phenotypic characteristics, geographic origin, and breeding history revealed
that MSY HTs were distributed unevenly (Figure 4B). Horse breeds influenced by
Arabians and Arabian studs carried HTs Ao and T, whereas all English
Thoroughbreds carried Tb, with Tb-dW approaching fixation. Tb and Tb-dW were
also predominant in European and American sport horses influenced by the English
Thoroughbred. The strong influence of Arabian (Ao) and Iberian (S and L)
lineages was evident in draft horses, pony breeds, and baroque breeds. The Ad
lineage was restricted to these three groups of breeds (Data S4).
MSY markers specific to particular
founder lines can be used to identify the origin of founder studs and to
determine their influence on the global horse population. We initially
reconstructed the paternal genealogy of descendants of three English
Thoroughbred founders, Darley Arabian (1700), Byerley Turk (1680), and
Godolphin Arabian (1724). Our samples included 110 descendants of Darley
Arabian, 22 of Byerley Turk, and seven of Godolphin Arabian. The pedigree
reconstruction and genotyping results are given in Figure S4. We assigned HT Tb-d to Darley
Arabian, and the predominant HT Tb-dW1 (defined by the variant rD) was in
agreement with the described HT3, representing the English Thoroughbred
stallion Whalebone (1807). Out of the five pedigree errors observed in the
descendants of Darley Arabian, one occurred in the lineage of King Fergus
(1775). Although fewer samples were available for the other two founder
stallions, we identified an association of HT Tb with Byerley Turk. The major
HT in the Godolphin Barb samples was Tb-g2, but based on the comparatively low
number of samples and their coalescence at Comus (1809) (Figure S4C), this founder stallion might
have had HT TB-g or even Tb. A reconstruction of the HTs of 32 other stallions
is presented in Data S4.
While the sub-branches of Tb (Tb-g,
Tb-k, Tb-r, Tb-dW, and Tb-dM) can definitely be attributed to English
Thoroughbred stallions, the basal HT Tb was also found in breeds with no
documented English Thoroughbred influence, such as the Hucul and Lipizzan
stallion lines. None of the classical breeds commonly used for refinement
(i.e., Spanish, Barb, and Arabian horses) carried Tb. To identify the origin of
Tb, we extended our samples by including the Akhal-Teke, the remnant of the Turkoman
horse, and found that Tb is the most frequent HT among 78 Akhal-Teke males
(81%, Figure 4B). Thus, Tb is likely of Turkoman
origin and spread widely by English Thoroughbred stallions. Additionally, the
presence of Tb in many European breeds with no documented influence of English
Thoroughbred stallions shows the influence of Turkoman stallions, independent
of the English Thoroughbred. This finding corresponds to the geopolitical
development of the region.
We conclude that the MSY crown group
(A-L-S-T) in present breeds is a footprint of the “Oriental horse,” with
haplogroup Tb attributed to a Turkoman origin and haplogroup Ao unambiguously
derived from “Original Arabians.” Some descendants of Original Arabians
clustered on the node basal to haplogroup T (Figure 4A, green circle). While this
finding seems to contradict our hypothesis that there were two distinct male
lineages for the Oriental founder populations, it might be explained by
incomplete lineage sorting in ancestral populations or by an underestimation of
private alleles. It might also simply reflect the inability of European horse
traders on the Oriental horse markets in the 19th century to
accurately identify “purebred” Arabian stallions. Variants specific to focal
lines will be key to further studies of high-resolution, unbiased horse
paternal genealogies. The horse non-repetitive MSY reference makes the
identification of biallelic markers for a larger number of samples
straightforward. It covers 1.67 Mb of the 15 Mb-spanning euchromatic part of
the horse MSY and consists only of MSY sequences that lack homology to the X
chromosome and to heterochromatic and repeat-rich regions. Based on the
estimated de novo mutation rate of the horse MSY and the length of
single-copy regions of the MSY, an average of 1 in 25 offspring of a particular
stallion are expected to carry a new mutation. Customized target enrichment
methods for MSY regions before sequencing make the identification of diagnostic
SNVs affordable. It is thus feasible to ascertain diagnostic SNVs for a broad
range of samples and stallion-line tracing, even for recently established
lines.
Although it was founded less than 1,000
years ago from only a few Oriental stallions, the A-L-S-T crown group now
accounts for most stallion lines in modern horse breeds. Similar to the
deep-branching northern lines, we expect additional rural breeds, especially
those from Asia, to harbor more autochthonous MSY variation. However, even with
more data from extant horses, MSY branches may be too shallow to reach back to
the beginning of domestication. Time series data from fossils should be
included in analyses to elucidate the domestication process from the male
perspective. The MSY phylogeny of modern horses can serve as a framework for
the interpretation of MSY diversity in autochthonous and ancient samples.
