1. Introduction
Canine coat color genetics is a subject of much interest to members of the dog breeding community, and the inherently visible and predominantly non-deleterious phenotypic presentation has encouraged many genetic research endeavors. Numerous commercial genotyping laboratories offer basic coat color assays. A recent publication documenting comprehensive color allele frequencies across breeds made note of an unexpected combination of alleles at the Agouti Signaling Protein (ASIP) gene [1]. The conventional understanding and nomenclature of ASIP inheritance consists of four characterized alleles, present in a dominance hierarchy: ay > aw > at > a. The ay allele is the most dominant and is responsible for fawn coloring, which consists of pheomelanin (yellow- or red-based pigment) at the base of a hair with a eumelanin (black- or brown-based pigment) at the tip [2]. On a molecular level, this allele is characterized by two neighboring amino acid substitutions, A82S and R83H, in exon 4 of ASIP [2]. The aw allele, which is the wild-type allele of ASIP, produces a wolf-sable banded hair pattern. When genotyping an individual, the aw allele is typically identified through a process of elimination when indicators for all other currently-tested ASIP alleles are absent [3]. This means that the aw “allele” is actually a haplotype representing the absence of any of the variant alleles at the ay, at, or a loci. The at allele is responsible for producing the tan points phenotype, with eumelanin on the torso of the dog and phaeomelanin on the face and extremities, and is due to a reverse SINE insertion at chr24:23,365,298–23,365,537 in intron 1 of ASIP [3]. The most recessive allele, a, produces a solid eumelanin color. The a allele is characterized by an R96C amino acid substitution in exon 4 of ASIP and occurs in conjunction with the SINE insertion found in at, thus the a allele has both the R96C mutation in exon 4 and the intron 1 SINE insertion [3,4]; these two variants are typically considered to be linked. The current practice of most commercial genetic testing laboratories offering coat color results is to do the following for ASIP for each dog: (1) test each variant locus (ay, at, and a) in ASIP, (2) identify which variant alleles are present at these three loci, (3) call aw as present in one copy when only one variant allele is detected at the three tested loci, or call aw as present in two copies when no variant alleles are detected, and (4) report a forced biallelic genotype, which is then generally interpreted according to the dominance hierarchy outlined above, with the most dominant allele present assumed to be dictating the phenotype. The recent publication of commercial laboratory-determined color allele frequencies, for which numerous authors overlap with those on this manuscript, noted the occurrence of what appeared to be three ASIP allele variants in multiple dogs [1]. At the time, this was interpreted to be the existence of the ay mutation and the at mutation occurring on the same chromosome, and was termed the ayt allele [1]. Subsequent genotype and phenotype data collection now allows us to present a more complete characterization of allele discrepancies at the ASIP locus in dogs, propose mechanisms for their occurrence, and outline the implications of such a finding. For the purpose of entertaining multiple possible explanations for the occurrence of these genotyping anomalies, we will herein refer to dogs genotyped with any combination of three or more currently-tested ASIP allele variants (ay, at, or a) as A3+. Using data generated via commercial laboratory coat color genotype-testing, with direct testing of published variants, our goal was to describe the existence of A3+ alleles, report which breeds they are observed in and the frequencies of these results by breed, and present observed phenotypes for A3+ dogs. Nomenclature as it currently appears in the published canine ASIP literature is used, and we suggest both future work and a shift in describing and analyzing the canine ASIP gene.
2. Materials and Methods 2.1. Datasets
Three datasets were independently considered for analysis (Table 1). The results of a prior study conducted by some of the current authors were utilized as published [1]; this study presented data as collected by Wisdom Health for their WISDOM PANEL analysis (Wisdom Health, Vancouver, WA, USA). In this initial study, referred to herein as Wisdom Health: True Colors (WHTC), the occurrence of unexpected combinations of ASIP allele variants (A3+) was observed in 679 dogs (out of 11,790) representing 14 breeds and the Dingo. This cohort has been well-described previously [1]. A new cohort consisting of 3189 purebred dogs representing 14 of these same breeds, plus three additional breeds not included in the initial study, was compiled and will be hence referred to as Wisdom Health: Agouti Expansion (WHAE). This cohort was selected based on breeds identified with the ayt allele from [1] and during routine data quality control at Wisdom Health. Some overlap in dogs does occur between WHTC and WHAE. The WHAE cohort was collected by Wisdom Health, Helsinki, Finland (formerly Genoscoper Laboratories) between July 2013 and September 2019. Most of the dogs in the WHAE cohort were from Finland (~one third) and the USA (~one quarter) with the remaining dogs representing a roughly global population. All Wisdom Health-generated data was collected and genotyped using the same protocols, as outlined below.
Other commercial laboratories provide canine ASIP allele genotyping, therefore, a supplemental cohort consisting of owner-reported ASIP genotypes, as produced by any one of six alternate (non-Wisdom Health) commercial laboratories, was assembled. This cohort contains 119 dogs representing three breeds and will be referred to as the Multi-Laboratory Agouti Supplement (MLAS). The supplemental cohort was genotyped using the proprietary protocols consistent with their respective commercial laboratories, and were reported as a biallelic genotype. Each of these laboratories has reported the occurrence of A3+ as an ayt allele in a forced biallelic genotype scenario. Phenotype data (coat color) was available for only a small number of dogs (n = 54) from the WHAE and MLAS cohorts. 2.2. Sample Collection and Genotyping For the WHTC and WHAE cohorts, all genetic analyses were carried out on DNA extracted from owner-collected, non-invasive buccal swab samples, or from blood/cheek swab samples collected at certified veterinary clinics in accordance with international standards for animal care and research. All dogs were owned as pets, and samples were voluntarily submitted by owners for commercial screening. All dog owners provided consent for use of their dog’s DNA for research purposes. Therefore, no additional Institutional Animal Care and Use Committee oversight is required. Where applicable, drawing of blood samples was approved by the Animal Ethics Committee of the State Provincial Office of Southern Finland, Hämeenlinna, Finland (permit number: ESAVI/6054/04.10.03/2012). Dogs were defined as purebred for the purposes of this study if registered with: Federation Cynologique Internationale, American Kennel Club, United Kennel Club, the UK Kennel Club, or an applicable single-breed registry for rare breeds.
DNA was extracted using standard protocols, and all dogs were analyzed on the WISDOM PANEL, MYDOGDNA, or OPTIMAL SELECTION platforms (Wisdom Health). Genotyping of all ASIP variants (ay, at, and a) was conducted on a custom-designed low-density Illumina Infinium bead chip using manufacturer-recommended protocols (Illumina, San Diego, CA, USA); the validation and genotyping quality control measures for this platform were previously described in detail [5,6]. A sample inclusion criteria call rate of 98% for all analyzed markers was enforced, and all genotype calls were manually curated. Bi-allelic genotypes for ASIP are obtained by testing for the occurrence of the ay A82S and/or R83H mutation (depending on the commercial laboratory; for the present study, all WHTC and WHAE dogs were genotyped for both alleles, and the alleles are confirmed to be in full linkage for all dogs in the WHAE cohort), the a R96C mutation, and the SINE insertion attributed to the at and a alleles. The absence of these mutations indicates the presence of the wild-type aw allele. For the purpose of this study, the raw genotypes at each of the mutations for all dogs in the WHTC and WHAE datasets were annotated separately and not transformed into the bi-allelic nomenclature. Genotypes were also obtained for the Melanocortin 1 Receptor (MC1R) [7,8] and Canine β-Defensin 103 (CBD103) [9,10] for phenotype analysis, as certain genotypes at these loci will mask expression of ASIP.
The individual ay and at allele genotypes detected as tri-allelic via the bead chip were confirmed in two dogs by standard capillary sequencing on an ABI3730 × 1 DNA Analyzer platform (ThermoFisher Scientific, Waltham, MA, USA) at the Sequencing Unit of the Finnish Institute of Molecular Medicine (Figure S1). Preparation and purification of PCR products for sequencing was carried out following manufacturer’s instructions, as previously described in detail [5,6]. Primer sequences are provided in Table S2, and all primers were verified to work on control (non-A3+) dogs.
3. Results 3.1. Detected Frequencies of A3+
The WHAE cohort consists of 3189 dogs, of which 158 dogs, representing 12 of the 17 breeds, were found to be A3+ (Table 1), a frequency of 4.95%. The proportion of individual dogs in each breed that genotyped as A3+ ranges from 1.32% (Whippet) to 63.64% (Dogo Argentino). The WHTC dataset had an A3+ frequency of 6.63% across 14 breeds plus Dingoes [1]. Five breeds (Berger Picard, Boston Terrier, Brussels Griffon, Great Dane, and Irish Terrier), identified in WHTC as consisting of at least one A3+ dog, were included in the WHAE dataset, and no additional incidents of A3+ dogs were identified. Conversely, three breeds (Basenji, East-Siberian Laika, and Kai Ken), in which A3+ was newly observed, were included in the WHAE dataset, but were not included in the WHTC dataset. Owner-reported genotypes obtained from any one of six other commercial testing laboratories (unaffiliated with the authors of this study) were recorded for 119 dogs (MLAS). Nineteen of these dogs were genotyped as A3+ (reported as an ayt allele), supporting that the detection of these anomalies is not dependent on laboratory-specific protocols, and is therefore not a simple genotyping error.
Observed allele combinations in WHAE consist of all potential ASIP bi-allelic pairs, with the exception of aw/a (Table 2). Three different combinations of three alleles were observed, and a single combination of four alleles was observed in seven separate dogs.
3.2. Phenotypic Impact of A3+
Color phenotypes were available for 54 A3+ dogs from the WHAE and MLAS cohorts, after the removal of dogs that were solid eumelanin due to the KB allele of CBD103 [9,10], solid phaeomelanin due to a homozygous e/e genotype at MC1R [8], or had white spotting patterns that effectively obscured the expressed ASIP phenotype. Four A3+ combinations were observed (Table 3), and some phenotypes were different than what would be expected under the existing allele hierarchy as described in the literature. If allele combinations and phenotypes are considered within breeds, multiple expression patterns begin to emerge. Thirty-two Tibetan Spaniels with known phenotypes have three allele variants each. Within this breed, the phenotype expressed is the one resultant of the allele that is present twice (Figure 1). For instance, a genotype of ay/ay/at produces the ay fawn phenotype, while a genotype of ay/at/at produces the at tan points phenotype. However, this pattern does not reflect the expression of alleles in other breeds, such as the Tibetan Mastiff. Four Tibetan Mastiffs with known phenotypes have been genotyped with three alleles, and in this breed both the ay/ay/at and ay/at/at combinations produce an ay fawn phenotype (Figure 1). Eleven East-Siberian Laikas were genotyped as having three or four ASIP alleles and also have available phenotypic information (Table 3). The allele combinations of ay/at/a, and ay/ay/at/at can produce phenotypes consistent with the aw wild-type allele (wolf sable), while the single East-Siberian Laika genotyped as ay/ay/at shows a fawn phenotype. In East-Siberian Laikas and across different breed backgrounds, the ay/at/at combination produced variable phenotypes, including the aw wild-type (wolf sable), ay fawn, and at tan points (Figure 1).
4. Discussion
The observation of more than two ASIP variants in a single dog presents an intriguing challenge within the small subset of breeds where it has been observed; these breeds will undoubtedly provide the key to ultimately determining what actual chromosomal rearrangement(s) have occurred. The breeds in which this scenario is observed range from molossoid flock guardians, to sight hounds, to companion breeds. Previous research suggests potential relatedness between some of these breeds, such as the Tibetan Spaniel and Tibetan Mastiff or the Maremma Sheepdog and Great Pyrenees, though these relationships point to distant common ancestral origins of multiple breeds rather than direct breed-to-breed progression [11,12]. Potential causes of the A3+ anomalies could be attributed to at least three different scenarios: (1) recombination within the ASIP gene, (2) duplication of part or all of the ASIP gene, or (3) an accepted ASIP allele is incorrectly attributed to and tested via an incompletely linked marker variant (Figure 2). When considering recombination as a potential explanation for the seeming appearance of multiple allele variants on the same chromosome, the positioning of the annotated allele variants must be considered. The SINE insertion in intron 1 is found in dogs with the tan point phenotype and, in conjunction with an R96C amino acid substitution in exon 4, in dogs with the recessive black phenotype. The ay allele, attributed to two adjacent amino acid substitutions at A82S R83H, is also in exon 4, thirteen codons upstream of the a allele variant. The genomic distance between the at SINE and ASIP exon 4 is approximately 27 kb. With an estimated recombination frequency of 1% per 1 cM, and a canine correction estimate of 1.55 cM/Mb [13], the frequency of recombination between the at SINE and exon 4 is 0.042%. We have observed dogs with A3+ genotypes involving the a allele, though these situations are such that they may consist of a normal a allele and a rearrangement only incorporating the other two alleles. Therefore, a recombination between the a and ay variants, situated at a distance of 39 bp apart, is not required for A3+ that includes an a allele. Under a recombination model, different breeds could display variable phenotype expression patterns of the A3+ allele variants.
An alternative theory to explain the detection of A3+ allele variants involves the partial or complete duplication of ASIP in a small number of dog breeds. The duplication, if non-functional, would allow for the detection of an allele variant without an observable impact on the expressed phenotype. In this way, assays designed to amplify genomic regions immediately surrounding the known variants would inadvertently identify the functional and non-functional gene copies. The duplication would be variable in terms of allele variants present, and also variable in the frequency of its occurrence, with the majority of dogs not possessing the duplication and genotyping as bi-allelic. Using the Tibetan Spaniel as an example (Figure 1), a dog with a tan point phenotype with a genotype of ay/at/at would have a functional ASIP genotype of at/at and a single duplication copy with the ay variant. Conversely, a fawn Tibetan Spaniel with a genotype of ay/ay/at would have a functional ASIP genotype of either ay/at or ay/ay, and the single duplication could be either ay or at, respectively. Some breeds that are observed to have the ay/ay/at combination are also observed to have a very low frequency of the at allele among dogs with a bi-allelic genotype. An example of this is the Whippet, which has a bi-allelic genotype allele frequency of 0.83% for the at allele. The three Whippets identified in this study that have an allele combination of ay/ay/at present a fawn phenotype (with brindle and/or white markings), consistent with the ay allele that is predominant in the breed. That would suggest, therefore, that the duplicated non-functional allele in this breed is the at.
Both the duplication and recombination theories could explain the variable inheritance patterns in different breeds and even the expression of the aw phenotype in the absence of a detectable aw allele. Under a recombination model with ayt on one chromosome, where the Illumina SNP genotyping methodology would indicate heterozygosity at both the ay and at loci, an aw haplotype could be present and not accurately detectable. If all alleles were functional, the dog should have a fawn (ay-dictated) phenotype, but would have an undetected aw chromosome that could be passed to offspring. Or, if the ay were non-functional, a dog could test as ay/at and still have an aw phenotype. Such scenarios must be considered in breeds with a known occurrence of the aw allele, such as the East-Siberian Laika. Conversely, a homozygous result at a variant locus (such as at/at) means that no wild-type allele is present at that locus when using the Illumina SNP genotyping methodology, and this indicates that the normal aw haplotype is not present. Intermediate coat color phenotypes could still be observed due to accumulation of ASIP mutations, including combinations of gain- or loss-of-function alleles, under a duplication model. A3+ scenarios are more difficult to elucidate phenotypically: in Figure 1F, the East-Siberian Laika presented has genotyped as ay/at/at, and presents a phenotype that could be consistent with a darkened form of ay fawn or the aw allele (haplotype), despite the fact that the homozygous at/at result means there is no wild-type allele at that locus (aw is not present). This ambiguity would allow for a possible intermediate phenotype of appearing more wolf sable (aw) due to the combination of accruing variant alleles, or appearing as a darkened fawn (ay/at), with the additional at allele present as a non-functional duplication.
The suggested duplication theory would require that a duplication event would have occurred multiple times in order for both the ay and at allele variants to be variably represented. A similar scenario of gene duplication with variable alleles has been observed in dogs before, specifically in regard to the production of the brindle phenotype. The brindle phenotype was shown to represent an allele of the CBD103 gene, kbr, dominant to ky and recessive to KB [9]. The molecular nature of kbr is a structural variant that affects gene copy number (personal communication, Drs. Barsh and Kaelin). Copy number variation of the ASIP gene has been reported to occur, and impact coat color, in domestic goats and sheep [14,15,16,17,18,19]. For these species, the copy number variant is purported to be functional, impacting pigmentation expression. A functional ASIP duplication in dogs, though as yet unsubstantiated, may also affect pigmentation and shading variation present in dogs with fawn, tan point, and wild-type (wolf sable) phenotypes. If any duplications have occurred, this increase in allele number for each of the tested variants could potentially create genotyping and/or test result interpretation errors. In reality, a recombination between the ay and at variants is not less likely than a scenario with multiple duplications. Mechanistic evaluations will be required to tease out these scenarios, always with the consideration that other coat color genes, both known and any yet undescribed, can influence the phenotype.
A third potential cause for the A3+ anomalies is the possibility that one or more of the identified allele variants of ASIP is actually a highly-linked marker and not a causal variant. We did not have phenotypes or family pedigree structure for most of the dogs in the present study, therefore we could not conduct association or linkage analyses in further investigations. While the exonic variants associated with ay and a, and the intronic SINE insertion attributed to at have been widely used commercially to predict inheritance and expression of coat colors and patterns, genotype discrepancies such as those detailed here could be caused by incomplete linkage between the assayed and causal variants. Alternate causal variants for the tan point phenotype have been proposed [20], though are not broadly incorporated into commercial testing panels.
These data highlight the necessity for additional investigation into the occurrence of multiple allele variants in relation to recombination events or the occurrence of ASIP gene duplications. For example, investigations could utilize DNA from dogs with known atypical genotypes in droplet digital PCR, long-range sequencing, and/or on commercially-available highly dense SNP arrays where SNP data can be accurately phased into haplotypes and used in linkage disequilibrium analyses and/or where raw SNP data intensity values can be analyzed for copy number changes. Accurate phenotypes and pedigrees would also facilitate teasing out haplotype-to-phenotype associations and linkage/recombination, not only on dogs with atypical ASIP results, but also across all dogs for verification. Any haplotype-to-phenotype analyses must, of course, keep other genes with epistatic functions in mind (e.g., MC1R, CBD103). Future work should also explore ASIP expression, which may help further decipher genotype/haplotype-to-phenotype correlations. The East-Siberian Laika, in particular, ought to have its ASIP gene sequence examined in its entirety, due to the incongruous genotype–phenotype observations.
With four already-published variants, including the wild type aw haplotype, in the canine ASIP gene, it is clear that this area of the genome is prone to change, and also clear that these changes are actively selected for in different breeds of domestic dog. Multiple additional species report point mutations, small insertions/deletions, and copy number variants of ASIP as causal for coat color phenotypes, suggesting great plasticity of this gene across species [18,19,21,22,23,24]. Therefore, it is reasonable to consider that either recombination or duplication, or both, could have occurred at this locus in dogs, possibly multiple times. It is probably wise to also consider that: (1) other alleles might exist at the known variants, (2) other variants might exist within the canine ASIP gene, and (3) linkage across the canine ASIP gene may not be 100%.
In addition, epigenetic mechanisms, such as methylation, could be in play. Variable coat color expression is seen in mice with a retrotransposon upstream of the agouti gene [25], where the range of color inversely correlates with DNA methylation states [26,27]. Such metastable epialleles, with variable phenotypes correlating directly with an epigenetic state, are not uncommon in mammalian genomes [28], making this yet another potential mechanism that must be considered in future investigations of canine ASIP expression.
Lastly, we have identified a situation that poses a challenge to commercial laboratories offering coat color genotyping, whereby a small portion of dogs will not correctly genotype for ASIP if the aw allele is masked by a recombination situation. The current data identifies six breeds for which this scenario definitely exists, given the determination of an aw allele through exclusion of at, ay, and a variants: Basenji, East-Siberian Laika, Great Pyrenees, Kai Ken, Lagotto Romagnolo, and Maremma Sheepdog. Under the same exclusion-based genotyping of aw, the previously published WHTC dataset adds the Anatolian Shepherd, Great Dane, Tibetan Mastiff, and Tibetan Spaniel breeds to this list, for a minimum of ten breeds that have some positive frequency of the aw allele and at least one observed incidence of greater than two identified ASIP variant alleles in a single dog [1]. Commercial laboratories should consider presenting a disclaimer with reported ASIP genotypes, particularly in breeds known to possess the aw allele (haplotype), that the reported presence or absence of that haplotype cannot be accurately detected in certain cases. In instances where a dog appears to present an aw phenotype in the absence of the appropriate genotype, it should be considered whether the breed in question may be affected in this context.
The current approach to understanding and interpreting the canine ASIP gene has worked well historically and is still valid in many breeds, but the data presented here indicate that it is not always accurate in several breeds. Adjustments are required, including changes in the typically-utilized terminology. Moving forward, it would be best to cease calling aw the wild-type “allele” and rather call it a haplotype, since it is actually the absence of the variant allele at any of the currently-tested loci on one chromosome. Canine coat color phenotypes, at least in known A3+ breeds, should be re-assessed via association with ASIP haplotypes, rather than defaulting to the existing allele hierarchy methodology. This is due to the evident segregation that is occurring between the alleles in these dogs, which is potentially causing a difference in phenotype compared to the existing literature and prevailing understanding in the dog genetic testing industry. Although A3+ dogs are rare, we were able to obtain phenotypes for 54 of 158 A3+ dogs in the present study, allowing us to speculate on the phenotype characterization. With this definitive report of A3+ dogs, laboratories providing coat color testing results should consider modifying their interpretation and reporting procedures. 5. Conclusions
We postulate that these unexpected A3+ genotypes have not been previously identified due primarily to their very low frequency across purebred dogs as a whole. The previous report indicated that greater than two ASIP alleles occurred in a single dog at a frequency of 0.41%, out of 11,790 dogs [1]. The recent increased accessibility and popularity of commercially-available canine genotyping panels has allowed for the data accumulation from sample sizes not previously seen in genetic research of this species, revealing these rare, though biologically relevant, atypical results. Significant additional experimental investigation and shifts in both nomenclature and analysis of the ASIP gene are required.
Enlarge this image.Figure 1. Conventional and A3+ ASIP genotype-phenotype combinations. (A-D) Tibetan Spaniels routinely express a tan point phenotype (A), produced by homozygous at alleles, and a fawn phenotype (B), produced with a dominant ay allele. When genotyped as A3+ (C,D), the allele that is present twice will dictate the phenotype. East-Siberian Laikas (E,F) possess all four known ASIP alleles naturally. The dominant ay allele produces a fawn phenotype (E), while an East-Siberian Laika genotyped as ay/at/at (F) could be ay fawn or aw wolf sable. In Tibetan Mastiffs (G) with ay/at/at, a fawn phenotype is expressed, a different pattern than that seen in Tibetan Spaniels (C).
Enlarge this image.Figure 2. The evolution of ASIP alleles with alternate means by which A3+ may have arisen. Blue dashed lines outline the possible duplication of the ay and at alleles into non-functional copies. Red dashed lines indicate the recombination of the ay and at alleles into a combined ayt allele. An example of a fawn colored Tibetan Mastiff and a tan-pointed Tibetan Spaniel, both having been genotyped as ay/at/at, are used to illustrate each mechanism.
| Breed | WHTC | WHAE | MLAS | |||||
|---|---|---|---|---|---|---|---|---|
| N | A3+ | % | n | A3+ | % | n | A3+ | |
| Anatolian Shepherd | 22 | 1 | 4.55% | 18 | 1 | 5.56% | - | - |
| Basenji | - | - | - | 74 | 1 | 1.35% | - | - |
| Berger Picard | 9 | 1 | 11.11% | 7 | 0 | 0.00% | - | - |
| Boston Terrier | 64 | 4 | 6.25% | 94 | 0 | 0.00% | 3 | 3 |
| Brussels Griffon | 44 | 1 | 2.27% | 44 | 0 | 0.00% | - | - |
| Dingo | 12 | 2 | 16.67% | - | - | - | - | - |
| Dogo Argentino | 9 | 6 | 66.67% | 11 | 7 | 63.64% | - | - |
| East-Siberian Laika | - | - | - | 109 | 15 | 13.76% | - | - |
| Great Dane | 79 | 1 | 1.27% | 163 | 0 | 0.00% | - | - |
| Great Pyrenees | 51 | 8 | 15.69% | 70 | 9 | 12.86% | - | - |
| Irish Terrier | 33 | 1 | 3.03% | 55 | 0 | 0.00% | - | - |
| Kai Ken | - | - | - | 19 | 1 | 5.26% | - | - |
| Lagotto Romagnolo | 116 | 8 | 6.90% | 1840 | 80 | 4.35% | - | - |
| Maremma Sheepdog | 18 | 3 | 16.67% | 24 | 5 | 20.83% | - | - |
| Spanish Greyhound | 25 | 1 | 4.00% | 72 | 1 | 1.39% | - | - |
| Tibetan Mastiff | 83 | 6 | 7.23% | 108 | 14 | 12.96% | 1 | 1 |
| Tibetan Spaniel | 46 | 1 | 2.17% | 177 | 20 | 11.30% | 115 | 15 |
| Whippet | 68 | 1 | 1.47% | 304 | 4 | 1.32% | - | - |
| Total | 679 | 45 | 3189 | 158 | 119 | 19 | ||
| No. of Non-Wild-Type Variants a | ay [2] | at [3] | a [4] | Interpreted Genotype b | No. of Observations (Out of 3189 Dogs) | Proportion of Genotypes |
|---|---|---|---|---|---|---|
| p.A82S | g.23365298ins239 | p.R96C | ||||
| 2 | 2/2 | 1/1 | 1/1 | ay/ay | 909 | 28.50% |
| 1 | 1/2 | 1/1 | 1/1 | ay/aw | 29 | 0.91% |
| 2 | 1/2 | 1/2 | 1/1 | ay/at | 376 | 11.79% |
| 2 | 1/2 | 1/2 | 1/2 | ay/a | 44 | 1.38% |
| 0 | 1/1 | 1/1 | 1/1 | aw/aw | 13 | 0.41% |
| 1 | 1/1 | 1/2 | 1/1 | aw/at | 62 | 1.94% |
| 1 | 1/1 | 1/2 | 1/2 | aw/a | 0 | 0.00% |
| 2 | 1/1 | 2/2 | 1/1 | at/at | 1303 | 40.86% |
| 2 | 1/1 | 2/2 | 1/2 | at/a | 265 | 8.31% |
| 2 | 1/1 | 2/2 | 2/2 | a/a | 30 | 0.94% |
| 3 | 2/2 | 1/2 | 1/1 | ay/ay/at | 41 | 1.29% |
| 3 | 1/2 | 2/2 | 1/1 | ay/at/at | 92 | 2.88% |
| 3 | 1/2 | 2/2 | 1/2 | ay/at/a | 18 | 0.56% |
| 4 | 2/2 | 2/2 | 1/1 | ay/ay/at/at | 7 | 0.22% |
a In the currently-utilized commercial genetic testing methodology, which determines the alleles present at the three ASIP variant loci, under the assumption that (1) aw is present in one or two copies when one or zero variant alleles are present, respectively, and (2) where a, when present, is in linkage with at. b Genotypes as they would be interpreted according to the current literature’s published hierarchy, and according to typical current commercial testing laboratory protocols.
| Alleles | Breed | n | Color | Phenotype Allele |
|---|---|---|---|---|
| ay/ay/at | Boston Terrier | 3 | fawn (with brindle) | ay |
| East-Siberian Laika | 1 | fawn | ay | |
| Great Pyrenees | 1 | fawn (with white) | ay | |
| Tibetan Mastiff | 1 | fawn | ay | |
| Tibetan Spaniel | 20 | fawn | ay | |
| Whippet | 3 | fawn (with brindle) | ay | |
| ay/at/at | East-Siberian Laika | 1 | wolf sable | aw |
| 2 | dark fawn or wolf sable | ay or aw | ||
| 2 | fawn | ay | ||
| 1 | tan points | at | ||
| Tibetan Mastiff | 3 | fawn | ay | |
| Tibetan Spaniel | 12 | tan points | at | |
| ay/at/a | East-Siberian Laika | 2 | wolf sable | aw |
| ay/ay/at/at | East-Siberian Laika | 1 | wolf sable | aw |
| Whippet | 1 | fawn (with brindle) | ay |
Supplementary Materials
The following are available online at https://www.mdpi.com/2073-4425/11/7/739/s1, Table S1: Overall frequencies of ASIP variants in each cohort. Frequencies calculated for WHTC and WHAE cohorts are based on the assumption that each dog has two alleles; this is likely not the case in reality, therefore these numbers are presumed to be slightly inaccurate. MLAS frequencies not provided because the sample recruitment was not random. Table S2: Primers used in verification Sanger sequencing. Table S3: Complete data ASIP genotypes for WHAE and MLAS cohorts. A1 = ancestral (wild-type) allele, A2 = derived (variant) allele. Figure S1: Sanger sequencing chromatograms of two A3+ dogs. Sanger sequencing was performed to confirm the Illumina SNP genotypes, and to verify that no nearby variants would have resulted in missing detection of an allele. Both dogs are A3+ Tibetan Spaniels, specifically, both dogs are homozygous for the at insertion and heterozygous for both ay missense variants, resulting in an interpreted genotype of ay/at/at. These were the only two samples from A3+ dogs with DNA available for additional sequencing. Primers (Table S2) were verified to work on control (non-A3+) samples.
Author Contributions
Conceptualization, D.L.D., H.A., J.D., A.D., A.M.H., and K.J.E.; methodology, D.L.D., H.A., J.D., and J.A.C.; validation, H.A. and J.D.; formal analysis, D.L.D., H.A., J.D., and J.A.C.; resources, A.M.H. and K.J.E.; data curation, D.L.D., H.A., J.A.C., and A.D.; writing-original draft preparation, D.L.D. and K.J.E.; writing-review and editing, D.L.D., H.A., J.D., J.A.C., A.M.H., and K.J.E.; supervision, D.L.D. and K.J.E.; project administration, K.J.E.; funding acquisition, J.D. and A.M.H. All authors have read and agreed to the published version of the manuscript.
Funding
Partial support for KJE was provided by the Office of the Director, National Institutes of Health (NIH) under award number K01-OD027051. This funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. This work was additionally made possible through the financial support of a commercially-affiliated funder, Wisdom Health, Vancouver, WA, USA as follows: several co-authors are employed by (H.A., J.D. and A.M.H.) or have previously consulted (K.J.E.) for this company. Wisdom Health and the employee co-authors had a significant role in providing raw data (data collection), assisting with the editing of the manuscript, and the decision to publish, but a lesser role in study design and data analysis. The specific roles of all authors are articulated in the 'author contributions' section.
Acknowledgments
Sincere appreciation is extended to Julia Mathlin of Wisdom Health, Helsinki, for laboratory assistance. The authors wish to express their heartfelt thanks to all owners that submitted DNA, genotypes, and phenotype data from their dogs for analysis.
Conflicts of Interest
Exist for H.A., J.D., and A.M.H. who are currently employed by Wisdom Health, and for K.J.E., who previously provided consulting services to Wisdom Health. The role of Wisdom Health, as partial funder for this study, is outlined in the 'funding' section. D.L.D., J.A.C., and A.D. declare no conflict of interest.
1. Dreger, D.L.; Hooser, B.N.; Hughes, A.M.; Ganesan, B.; Donner, J.; Anderson, H.; Holtvoigt, L.; Ekenstedt, K.J. True Colors: Commercially-acquired morphological genotypes reveal hidden allele variation among dog breeds, informing both trait ancestry and breed potential. PLoS ONE 2019, 14, e0223995.
2. Berryere, T.G.; Kerns, J.A.; Barsh, G.S.; Schmutz, S.M. Association of an Agouti allele with fawn or sable coat color in domestic dogs. Mamm. Genome 2005, 16, 262-272.
3. Dreger, D.L.; Schmutz, S.M. A SINE insertion causes the black-and-tan and saddle tan phenotypes in domestic dogs. J. Hered. 2011, 102, S11-S18.
4. Kerns, J.A.; Newton, J.; Berryere, T.G.; Rubin, E.M.; Cheng, J.F.; Schmutz, S.M.; Barsh, G.S. Characterization of the dog Agouti gene and a nonagouti mutation in German Shepherd Dogs. Mamm. Genome 2004, 15, 798-808.
5. Donner, J.; Kaukonen, M.; Anderson, H.; Möller, F.; Kyöstilä, K.; Sankari, S.; Hytönen, M.; Giger, U.; Lohi, H. Genetic panel screening of nearly 100 mutations reveals new insights into the breed distribution of risk variants for canine hereditary disorders. PLoS ONE 2016, 11, 1-18.
6. Donner, J.; Anderson, H.; Davison, S.; Hughes, A.M.; Bouirmane, J.; Lindqvist, J.; Lytle, K.M.; Ganesan, B.; Kaukonen, M.; Forman, O.P.; et al. Frequency and distribution of 152 genetic disease variants in over 100,000 mixed breed and purebred dogs. PLoS Genet. 2018, 14, 1-20.
7. Everts, R.E.; Rothuizen, J.; Van Oost, B.A. Identification of a premature stop codon in the melanocyte-stimulating hormone receptor gene (MC1R) in Labrador and Golden retrievers with yellow coat colour. Anim. Genet. 2000, 31, 194-199.
8. Newton, J.M.; Wilkie, A.L.; He, L.; Jordan, S.A.; Metallinos, D.L.; Holmes, N.G.; Jackson, I.J.; Barsh, G.S. Melanocortin 1 receptor variation in the domestic dog. Mamm. Genome 2000, 11, 24-30.
9. Kerns, J.A.; Cargill, E.J.; Clark, L.A.; Candille, S.I.; Berryere, T.G.; Olivier, M.; Lust, G.; Todhunter, R.J.; Schmutz, S.M.; Murphy, K.E.; et al. Linkage and segregation analysis of black and brindle coat color in domestic dogs. Genetics 2007, 176, 1679-1689.
10. Candille, S.I.; Kaelin, C.B.; Cattanach, B.M.; Yu, B.; Thompson, D.A.; Nix, M.A.; Kerns, J.A.; Schmutz, S.M.; Millhauser, G.L.; Barsh, G.S. A β-defensin mutation causes black coat color in domestic dogs. Science 2007, 318, 1418-1423.
11. Parker, H.G.; Dreger, D.L.; Rimbault, M.; Davis, B.W.; Mullen, A.B.; Carpintero-Ramirez, G.; Ostrander, E.A. Genomic analyses reveal the influence of geographic origin, migration, and hybridization on modern dog breed development. Cell Rep. 2017, 19, 697-708.
12. Talenti, A.; Dreger, D.L.; Frattini, S.; Polli, M.; Marelli, S.; Harris, A.C.; Liotta, L.; Cocco, R.; Hogan, A.N.; Bigi, D.; et al. Studies of modern Italian dog populations reveal multiple patterns for domestic breed evolution. Ecol. Evol. 2017, 8, 2911-2925.
13. Dumont, B.L.; Payseur, B.A. Evolution of the genomic rate of recombination in mammals. Evolution 2008, 62, 276-294.
14. Han, J.L.; Yang, M.; Yue, Y.J.; Guo, T.T.; Liu, J.B.; Niu, C.E.; Yang, B.H. Analysis of agouti signaling protein (ASIP) gene polymorphisms and association with coat color in Tibetan sheep (Ovis aries). Genet. Mol. Res. 2015, 14, 1200-1209.
15. Rochus, C.M.; Westberg Sunesson, K.; Jonas, E.; Mikko, S.; Johansson, A.M. Mutations in ASIP and MC1R: Dominant black and recessive black alleles segregate in native Swedish sheep populations. Anim. Genet. 2019, 50, 712-717.
16. Henkel, J.; Saif, R.; Jagannathan, V.; Schmocker, C.; Zeindler, F.; Bangerter, E.; Herren, U.; Posantzis, D.; Bulut, Z.; Ammann, P.; et al. Selection signatures in goats reveal copy number variants underlying breed-defining coat color phenotypes. PLoS Genet. 2019, 15, e1008536.
17. Liu, M.; Zhou, Y.; Rosen, B.D.; Van Tassell, C.P.; Stella, A.; Tosser-Klopp, G.; Rupp, R.; Paljiere, I.; Colli, L.; Sayre, B.; et al. Diversity of copy number variation in the worldwide goat population. Heredity 2019, 122, 636-646.
18. Fontanesi, L.; Beretti, F.; Riggio, V.; Gomez Gonzalez, E.; Dall'Olio, S.; Davoli, R.; Russo, V.; Portolano, B. Copy number variation and missense mutations of the agouti signaling protein (ASIP) gene in goat breeds with different coat colors. Cytogenet. Genome Res. 2009, 126, 333-347.
19. Norris, B.J.; Whan, V.A. A gene duplication affecting expression of the ovine ASIP gene is responsible for white and black sheep. Genome Res. 2008, 18, 1282-1293.
20. Ciampolini, R.; Cecchi, F.; Spaterna, A.; Bramante, A.; Bardet, S.M.; Oulmouden, A. Characterization of different 5'-untranslated exons of the ASIP gene in black-and-tan Doberman Pinscher and brindle Boxer dogs. Anim. Genet. 2013, 44, 114-117.
21. Rieder, S.; Taourit, S.; Mariat, D.; Langlois, B.; Guerin, G. Mutations in the agouti (ASIP), the extension (MC1R), and the brown (TYRP1) loci and their association to coat color phenotypes in horses (Equus caballus). Mamm. Genome 2001, 12, 450-455.
22. Eizirik, E.; Yuhki, N.; Johnson, W.E.; Menotti-Raymond, M.; Hannah, S.S.; O'Brien, S.J. Molecular genetics and evolution of melanism in the cat family. Curr. Biol. 2003, 13, 448-453.
23. Hiragaki, T.; Inoue-Murayama, M.; Miwa, M.; Fujiwara, A.; Mizutani, M.; Minvielle, F.; Ito, S. Recessive black is allelic to the yellow plumage locus in Japanese quail and associated with a frameshift deletion in the ASIP gene. Genetics 2008, 178, 771-775.
24. Chandramohan, B.; Renieri, C.; La Manna, V.; La Terza, A. The alpaca agouti gene: Genomic locus, transcripts and causative mutations of eumelanic and phaeomelanic coat color. Gene 2013, 521, 303-310.
25. Oey, H.; Isbel, L.; Hickey, P.; Ebaid, B.; Whitelaw, E. Genetic and epigenetic variation among inbred mouse littermates: Identification of inter-individual differentially methylated regions. Epigenetics Chromatin 2015, 8, 1-12.
26. Waterland, R.A.; Jirtle, R.L. Transposable elements: Targets for early ntritional effects on epigenetic gene regulation. Mol. Cell Biol. 2003, 23, 5293-5300.
27. Blewitt, M.E.; Vickaryous, N.K.; Paldi, A.; Koseki, H.; Whitelaw, E. Dynamic reprogramming of DNA methylation at an epigenetically sensitive allele in mice. PLoS Genet 2006, 2, 1-7.
28. Ekram, M.B.; Kang, K.; Kim, H.; Kim, J. Retrotransposons as a major source of epigenetic variations in the mammalian genome. Epigenetics 2012, 7, 370-382.
Dayna L. Dreger1, Heidi Anderson2, Jonas Donner2, Jessica A. Clark1, Arlene Dykstra3, Angela M. Hughes4 and Kari J. Ekenstedt1,*
1Department of Basic Medical Sciences, College of Veterinary Medicine, Purdue University, West Lafayette, IN 47906, USA
2Wisdom Health, FI-00290 Helsinki, Finland
3Coronado, Sturgeon County, AB T0A 1N4, Canada
4Wisdom Health, Vancouver, WA 98683, USA
*Author to whom correspondence should be addressed.
© 2020. This work is licensed under http://creativecommons.org/licenses/by/3.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Details
; Anderson, Heidi; Donner, Jonas
; Clark, Jessica A
; Dykstra, Arlene; Hughes, Angela M; Ekenstedt, Kari J