The Kelpie breed is an iconic dog breed developed in Australia in the late 19th century. The Working Kelpie Council of Australia reports that the breed was developed from Scottish Smooth Collie and Farm Collie stock, and was founded upon a small number of key individuals. The breed is renowned for its resilience, working in harsh (hot, dry, and prickly) conditions in the Australian outback [1
]. Since its inception the Kelpie population has become divided, as particular lines were developed for exceptional performance in specialist aspects of working dog performance. In addition, one group from the population was separated to participate in dog conformation showing. This breed division has resulted in readily visible phenotypic differences between the groups. For instance, there are marked coat color and patterning frequency differences between the working and conformation Kelpie types.
The Australian Kelpie (AK) (Figure 1
c) that has been bred for conformation and registered with the Australian National Kennel Council most often has a solid/self-color coat in either black or brown [1
]. Homozygous coding variants and compound heterozygous genotypes coded by the gene 5
,6-dihydroxyindole-2-carboxylic acid oxidase precursor
) may result in a brown color [3
]. Solid black (self-color) may be coded by either a recessive variant at the Agouti signaling protein
) and RNA-Binding Protein (Autoantigenic, Heterogeneous Nuclear Ribonucleoprotein-Associated With Lethal Yellow)
) loci or a dominant variant at the canine β
) locus [5
]. Dogs that are genetically brown as a result of recessive alleles at TYRP1
are observed to have varying pigment intensity, and the genes underlying this difference in hue are thus far undescribed in dogs with normal coat variation (as opposed to those affected by albinism) [4
Australian Working Kelpies (AWK) (Figure 1
b) registered with the Working Kelpie Council of Australia are bred for their livestock herding capability. Dogs of this variety are frequently taller and have longer bodies and muzzles than the AK type. These dogs typically exhibit the same two main base colors (black and brown), but frequently exhibit tan markings on the black or brown coat. Confusingly, dogs that are brown due to the activity of TYRP1
are colloquially referred to as red in the breed. The tan markings are predicted to be due to alleles at either the ASIP
locus (A-locus) or the CBD103
locus (K-locus) [5
]. The occurrence of the brindle phenotype encoded by CBD103
] is undescribed in either variety of the Kelpie breed.
Other coat colors that exist in AWKs include yellow (known as ginger
in the breed), which can be driven by the pheomelanin variant at Melanocortin-1 receptor
]. A red or yellow coat color may also be caused by the dominant yellow variant (Ay
) at ASIP
]. Blue, fawn, and cream (all resulting from dilution at Melanophilin
) of black, brown, and ginger, respectively) [9
] are observed in AWKs, and these may occur with or without tan markings.
White markings are rarely observed in the breed, and when present, typically result from recent out-crossing with other livestock herding breeds, such as the Border Collie. The white patterning in most dog breeds is controlled by variants in the melanocyte-specific promoter region of Micropthalmia transcription factor
), which in combination impact the extent of white markings in the dog. The first variant in the region that is expected to impact the extent of white is a short-interspersed nuclear element (SINE) insertion three kilobases upstream of the melanocyte-specific promoter for the gene. The second variant associated with the extent of white is a length polymorphism in the immediate vicinity of the melanocyte-specific first exon [11
]. In general, dogs homozygous for the SINE element are expected to have either piebald or extreme white markings. At the length polymorphism, dogs with longer alleles are expected to have some white markings. The alleles behave in a co-dominant manner, with intermediate white markings observed on heterozygotes of the variant allele types.
Throughout the history of the Kelpie breed, speculation has persisted that the Australian wild dog Canis dingo
(Dingo) contributed to its founding. Such ideas were likely fueled by statements attesting this by respected individuals such as geneticist Dr. R. B. Kelley, the author of a widely recognized book used in breeding and training livestock herding dogs first published in 1942 [12
]; more recently, others suggest, albeit without peer review, that DNA demonstrates common ancestry with the Dingo [13
]. Others who claim personal knowledge of the breed founding lines [14
], or who report that attempted crosses between Kelpies and Dingoes to improve working traits failed regardless of the high quality of the Kelpies used in the crossing [15
], offer evidence to the contrary.
The Dingo (Figure 1
a) has an external phenotype that is quite like the Kelpie. The two dog types are of similar size (the Dingo may be slightly larger). Both have erect ears. While Dingoes are most commonly yellow-gold in color as adults, they can occur with dark coat colors on occasion. Both dog types have similar hair length and texture. Many laypeople believe that the ginger and cream variants in the Kelpie breed (Figure 1
d) are derived from the Dingo ancestral type. In common with the Dingo, both breed varieties most commonly exhibit erect ears (pricked as opposed to drop-ears). While ears with complete drop are very rare in the Kelpie, some individuals have ears that are incompletely erect. Segregation of single nucleotide polymorphism (SNP) markers with the ear phenotype has been considered extensively in the literature, although no functional mutation has been proposed [16
]. The gene Methionine sulfoxide reductase 3
) has been suggested as a strong regional candidate gene for the phenotype [17
Our analysis will determine the mutational basis of commonly observed external morphologies (coat color and ear phenotype) in the two Kelpie breed varieties at loci for which variants have been described. We include a description of two insertion variants in the 3′ untranslated region (UTR) of the gene MSRB3 that have not been reported in the literature. Both longer variants in the MSRB3 UTR are in near complete linkage disequilibrium, with a potentially functional array SNP that has been strongly associated with prick and drop ears in many breeds. None of the analyzed variants support co-ancestry with the Australian Dingo.
Despite the commonly held view that the Dingo is an ancestral contributor to the Australian Kelpie [12
], there is scant evidence in the current sequencing data to support this assertion. It is predominantly the morphologic similarities between the Dingo and Kelpie varieties that underlie the conjecture of introgression of Dingo into the original Kelpies. Here, we observe that different alleles segregate in the domestic and wild canids at the loci underpinning the same external morphologies that on the surface appear to be shared between the Dingo and the two Kelpie varieties.
The A-loci (ASIP-RALY on chromosome 24) underlie solid color (black/brown/yellow) and saddle or tan-point coat markings. At the ASIP-locus, most wolves would have the wild-type allele, whilst dingoes and many domestic dogs have other variants. Kelpies and Dingoes are strongly diverged at the A-loci (ASIP-RALY on chromosome 24), where in Dingoes the dominant yellow (Ay) haplotype has the highest frequency in conjunction with a variant at the nearby RALY gene, supporting a “saddle” pattern of markings. At these loci, the Kelpie exclusively has a haplotype that supports the non-agouti phenotype, with tan-point markings (at) at the RALY locus and wild-type variants in ASIP. While the saddle pattern from RALY is not readily visible in the Dingo, this is likely affected by the Ay haplotype at ASIP.
Ginger/yellow in the Kelpie is likely caused by the recessive CFA5 g.63694334G>A variant coded by the gene MC1R
, as this allele was observed in the heterozygous state in 3 of 12 sequenced Kelpies (both varieties). In the Dingo, yellow coat color is driven by the Ay
allele at the ASIP
locus and homozygosity for the insertion variant at CBD103
(reference allele), which enables colors encoded by variants at the A-loci (RALY-ASIP
) and E-locus (MC1R
) to be observed [5
]. Among 26 Dingoes (including 25 Dingoes sequenced previously for the MC1R
coding region), only one was heterozygous for the MC1R
CFA5 g.63694334G>A variant (allele frequency <2%) [20
CBD103 dominant black
(del/-) is perhaps better described as self-color, as it is epistatic to the genotype at ASIP
, where the phenotype is black in the presence of homozygosity for the non-agouti allele (most usually at
) at the A-locus [5
]. The representative sequenced Dingo in this study was homozygous for the insertion (pheomelanin expressing) allele of this gene. Among 25 Dingoes sequenced for the coding portion of this gene, all were homozygous for the insertion variant, but one was equivocal and possibly heterozygous [20
Among Kelpies, the major coat color difference between the two assessed varieties is the presence or absence of tan markings. We show that this variation is controlled by the CBD103
gene (commonly known as the K-locus) [5
]. Both Kelpie breed varieties are concordant (homozygous wild-type) at ASIP
, which has been proposed in the past to control black versus black and tan. Both Kelpie varieties actively segregate TYRP1
-red, although ginger/cream are disallowed by the Australian National Kennel Council breed standard, which affects the AK variety. Melanistic mask was observed to freely segregate in the AWK. The only observed coding variant at TYRP1
is p.Q331X, and this is the same coding variant for brown coat color for both AK and AWK groups. Differences in pigment intensity or hue in brown kelpies remain unexplained by the loci described thus far. The Dingo RKW13760 was homozygous for the black eumelanin expressing variant at TYRP1
Dilution as coded by MLPH was not assessed thoroughly in this analysis, due to poor sequence quality caused by high local G–C nucleotide content. One dog with low-quality cover (AWK USCF640) had a single sequencing read that is concordant with the dog being heterozygous for the presence of the MLPH splicing variant. Dilute animals are known to be present within the variety.
The length polymorphism in the melanocyte-specific promoter of MITF
defines a spectrum of solid to spotted coat markings in dogs. The allele that is most commonly observed in both Kelpie varieties is yet undescribed in the literature (length of 33 bases, coded by 10C9A2G12A). The absence of the SINE element CFA20 g.21836232_21836427ins>del at the MITF
locus would suggest that if white markings were to occur in either Kelpie variety, they would be of modest extent. The allele observed at MITF
in the Dingo has been previously described as wild-type
35B (observed in a Scandinavian wolf) [30
]. The sequenced Dingo was additionally homozygous for the SINE element insertion at CFA20 g.21836232_21836427, which is predictive of the presence of piebald or extreme white markings. The SINE element insertion and others of the longer haplotypes at CFA20 g.21839321_21839366 (indel) have been commented upon [30
], because the lengths of the alleles would predict the presence of white markings, which are rare in wolves. However, it is not uncommon for Dingoes to have white paws and chest. It is possible that alleles at MITF
are epistatic with those at the A-locus, to influence the penetrance of white markings as coded by the MITF
Erect ears are common between the Dingo and both Kelpie varieties. The SNP variant [17
] detected in the Kelpie at the ear locus concurs with that of the Dingo; however, the insertion allele in the nearby MSRB3_UTR (this study) differs between the domestic and wild dogs, suggesting that the variants have no recent co-ancestry. Other aspects of ear structure than ear erectness are as yet unpublished in the canine scientific literature.
This study is limited by the genetics observable in modern-day Kelpies. It is possible that attempts at introgression of the Dingo into the Kelpie occurred historically, but that the Dingo genes did not persist in the population at the loci observed. The possible presence of the dominant black CBD103 allele in one sequenced Dingo may suggest introgression in that canid, or may represent a natural low frequency variant in that population. In either case, that animal would be expected to have a black coloration. Further investigation of the MC1R sequence for this animal (Alpine5) reveals a likely coding variant near the start codon for the gene, which may result in a gold coloration for this animal (Accession KF586907.1).