Capturing goats: documenting two hundred years of mitochondrial DNA diversity among goat populations from Britain and Ireland

1. Background

The domestic goat (Capra hircus) is an agricultural species of global importance, prized in regions of marginal pasture [1,2]. Unfortunately, the advent of intensive selection schemes seeking to create ‘improved’ domestic breeds, coupled with high extinction rates among wild ibex and feral goat populations, has begun to deplete the bank of genetic diversity available to breeders [3,4]. Surveys of the genetic variation in surviving feral goat populations may help to identify diverse and irreplaceable groups in need of conservation efforts. The indigenous goat populations of Britain and Ireland, a proposed Neolithic introduction [5], are hypothesized to be one such candidate. These landrace animals, collectively described as ‘Old Goats’, were predominantly used in agriculture up until the late nineteenth century when they began to be replaced by continental breeds [6]. Today they survive only in small feral herds, whose existence, in recent decades, has been threatened by habitat loss, culling and the ongoing impact of continental introgression.

Natural history collections house an immense bank of taxidermy specimens that represent an extensive, although relatively untapped, source of genetic information [7–9]. In recent years, however, these collections have been exploited in a number of landmark studies, which have documented recent temporal changes in diversity and population structure in a range of species [7,10–12], by harnessing the power of high throughput sequencing (HTS). These genetic resources are likely to be especially important in the case of domestic animals, which have seen dramatic drops in effective population sizes and extensive transportation over the past two centuries owing to the advent of industrialized breeding [3,13].

Here, we present a genetic survey of 15 historical samples, derived from taxidermy specimens, spanning roughly 200 years in Britain and Ireland, as well as nine modern samples from dairy and feral populations in Ireland. These are analysed using whole mitochondrial genomes generated through in-solution capture and HTS.

2. Material and methods

DNA extraction and Illumina library preparation were performed using standard protocols (electronic supplementary material). Samples were subsequently subjected to target-enrichment using an Agilent custom-made mtDNA in-solution (RNA) capture (electronic supplementary material). Sequencing was performed on the Illumina MiSeq platform (50 bp single-end).

NGS reads were aligned and filtered following standard pipelines (electronic supplementary material). Fasta consensus sequences [14] were called using UGENE [15].

Full mitochondrial genomes for 53 European and Turkish goats (electronic supplementary material, table S1) [16] were used alongside 23 whole mitochondrial sequences of sufficient coverage from this study (greater than 85% of sequence covered at greater than or equal to 3X) to construct a maximum-likelihood phylogeny. The HKY85 [17] substitution model was selected for tree-building using jmodeltest2 [18,19], and performed with 1000 bootstrap repeats.

HV1 control regions (482 bp) of 132 western and central European domestic goats [20] were retrieved from GenBank (electronic supplementary material, table S2). These samples were collected from indigenous herds belonging to remote villages and thus make an ideal comparative dataset. Notably, this included sequences from four additional ‘Old Irish’ goats, two English Kielder goats and four Welsh Bagot goats. These data, alongside 22 HV1 control regions from the current study, were used in the creation of a maximum-likelihood phylogeny, constructed using the Kimura two-parameter model of substitution (1000 bootstrap repeats), as has been standard in previous studies of goat D-loop variation [20,21]. Additionally, a median-joining [22] haplotype network of the same dataset was created using PopART (http://popart.otago.ac.nz).

Multiple sequence alignments were performed with Muscle [23]. Alignments and maximum-likelihood phylogeny constructions were implemented in seaview v. 4.5.4 [24].

3. Results and discussion

In total 24 whole mtDNA genomes were generated at an average coverage of 40X (table 1; electronic supplementary material, figure S1). As expected, modern samples yielded higher numbers of total sequenced reads compared with the majority of historical samples. However, historical samples in general had a higher capture specificity (electronic supplementary material, figure S2). This resulted in a higher average coverage recovered for historical samples (44X) rather than modern samples (33X), albeit with a greater variance (4X–97X) than moderns (12X–51X). These data reiterate recent findings suggesting taxidermy samples are an excellent choice for study in population genetic experiments [10], and demonstrate the effectiveness of the mitochondrial RNA capture method for these types of specimen.

The whole mitochondrion phylogeny (figure 1) showed the vast majority of historical samples from this study to fall into two groupings at high bootstrap probability (81 and 87, respectively), both of which are composed solely of samples from Britain and Ireland. Notably, all modern Irish dairy goats place outside of these groupings. Furthermore, two modern feral goats from the Burren, Co. Clare (M-Irish3 and M-Irish4), both of whom showed morphological signs of Swiss introgression (e.g. chin tassels, large ears, conformation) [25], also fall away from the historical clades, highlighting the impact mass importation of continental breeds [6] has had on genetic variation in both commercial and feral Irish goat herds. Importantly from a conservation perspective, two modern goats (M-Irish1 and M-Irish2), belonging to a feral herd in Mulranny, Co. Mayo, are seen clustering with high bootstrap support alongside a clade of three historical ‘Old Scottish’ goats (H-Scot1, H-Scot2 and H-Scot8), two of which originate from an extinct population in the Isle of Skye that existed before 1900.

• Download figure
• Open in new tab
• Download PowerPoint

Owing to the limited sample size of the whole mitochondrion tree, a HV1 control region phylogeny was constructed to investigate whether the above clades remained consistent within a larger western and central European dataset (figure 2a). Bootstrap values (not shown) were uniformly low across the haplogroup A branches of the phylogeny, reflecting a lack of diversity between sequences. This is illustrated by the median-joining network in figure 2b, where many haplotypic groupings are defined by only one or two mutations, emphasizing the need for whole mitochondrion sequences in studies of goat maternal diversity. Despite the lack of bootstrap support, the HV1 control region phylogeny, taken together with the network, support the conclusions drawn from the whole mitochondrion tree. Two peripheral groupings, dominated by historical British and Irish samples are seen apart from the remainder of European haplogroup A variation, with branching patterns similar to those seen in the whole mitochondrion phylogeny. This includes the placement of the Mulranny goats with those from the Isle of Skye, alongside a third previously published Irish ‘Old Goat’ sample [20]. Interestingly, this network branch is also seen to contain a Norwegian landrace and native Polish goat.

• Download figure
• Open in new tab
• Download PowerPoint

Overall, these results would suggest some degree of mitochondrial geographical structure between the ‘Old Goats’ of insular Europe and their continental counterparts, an expected characteristic of island populations. Apart from the branchings described above and a few other possible exceptions, most notably a cluster containing only Welsh, English and Icelandic goats alongside M-Improv1, the HV1 tree and network both show decidedly low levels of geographical structure. This has been previously observed among goat populations, postulated to be the outcome of extensive movement and trade of goats in their history [21].

4. Conclusion and perspectives

This to our knowledge is the first study of livestock genetics to apply HTS to taxidermy specimens and it further demonstrates the high rates of efficiency of in-solution hybrid capture as a target-enrichment method for skin samples [26]. This cost efficient approach can give access to the wealth of genetic information housed in taxidermic collections, an invaluable resource when studying the population history of domesticates. Extensive transportation and intensive selection pressure over the past two centuries has likely obscured more ancient patterns of genetic variation within these species and dramatically reduced their levels of genetic diversity. The study of museum specimens also allows for the identification of present day, unique populations that have managed to retain some of this past diversity, many of which are being pushed to the brink of extinction today. We demonstrate this potential by drawing a link between extinct ‘Old Goat’ populations living on the Isle of Skye and endangered herds living in Mulranny, Co. Mayo today. Our results would also seem to support documented morphological observations suggesting feral herds in the Burren have been subject to a larger degree of introgression in comparison to the Mulranny population [25]. Overall, these findings identify Mulranny goats as some of the last modern representatives of the ‘Old Irish’ type of goat, once ubiquitous throughout the island, and highlight them for much needed conservation efforts.

Ethics

All animal research was carried out in compliance with standards outlined in EU legislation. Humane, non-invasive sampling (hair follicles) was carried out on all modern goats, with the exception of two previously deceased Mulranny individuals from whom skin was obtained.

Data accessibility

Fasta files of whole mitochondrion consensus sequences are available in the NCBI GenBank (KY564246-KY564269) and Dryad (http://dx.doi.org/10.5061/dryad.g9v24) [14].

Authors' contributions

L.M.C., M.D.T. and V.M. produced and analysed the genetic data. E.K.F. designed the RNA capture. S.C., R.E. and R.W. provided samples. D.G.B. supervised this study and contributed to its design. L.M.C, M.D.T. and V.M. wrote and edited the manuscript with contributions from all authors. All authors approved the final version and agree to be accountable for all aspects of this work.

Competing interests

S.C., R.W. and R.E. are or have been affiliated with the Old Irish Goat Society.

Funding

Research was funded by ERC Investigator Grant 295729-CodeX to D.G.B. L.M.C. is funded by the Irish Research Council (GOIPG/2013/1219). The European Union's Seventh Framework Programme (FP7/2010-2014) provided funding for this project under grant agreement no. 244356 – “NextGen.”

Footnotes

Electronic supplementary material is available online at https://dx.doi.org/10.6084/m9.figshare.c.3693208.