Correction: Mutations in MITF and PAX3 Cause "Splashed White" and Other White Spotting Phenotypes in Horses.

[This corrects the article DOI: 10.1371/journal.pgen.1002653.].

GWAS Identifies a Region Containing the SALL1 Gene in Variation of Pigmentation Intensity Within the Chestnut Coat Color of Horses.

Chestnut coat color in horses is determined by a missense mutation within the MC1R gene. However, the intensity of the chestnut color can vary widely within individuals possessing this genotype. Here, we investigated this variation using standardized photographs of 96 horses. Each horse was ranked lightest to darkest within the cohort for phenotype by 3 blinded observers. A genome-wide association study utilizing the relative shade ranking as the phenotype and using 268 487 single-nucleotide polymorphisms (SNPs) genotyped using the Affymetrix Equine 670k array identified a single significantly associated region on chromosome 3 (P = 2.934 × 10-8). Analysis of whole-genome sequences for horses spanning the diverse range of chestnut color identified candidate SNPs within the coding sequence of the only gene in the region: SALL1. The function of SALL1 is largely unknown, though it is predicted to interact with the Hermansky-Pudlak Syndrome type 1 (HPS1) protein, which causes partial albinism in humans. However, with only one study suggesting a circumstantial influence of the SALL1 protein on pigmentation, additional work is needed to confirm this new coat color locus in larger populations and investigate the function of this protein for impacts on equine health.

Two Variants of KIT Causing White Patterning in Stock-Type Horses.

Over 30 polymorphisms in the KIT Proto-Oncogene Receptor Tyrosine Kinase (KIT) gene have been implicated in white spotting patterns ranging from small areas to full dermal depigmentation in the horse. We performed a candidate-gene exon sequencing approach on KIT and MITF, 2 known causatives of white spotting patterns, within 2 families of horses of unknown white spotting. Family 1 (Fam1, N = 5) consisted of a Quarter Horse stallion and 4 offspring with white spotting pattern ranging from legs, lower ventral, and head regions with jagged borders, to almost complete white. The second family (Fam2, N = 7) consisted of 6 half-sibling American Paint Horse/Quarter Horse and their dam, demonstrating unpigmented limbs with belly spots and an extensive white patterning on the face. This approach resulted in 2 variants significantly associated with familial phenotypes, where Fam1 variant is an indel leading to a frameshift mutation, and Fam2 a non-synonymous SNP. We validated the variants within an unrelated population of horses (Fam2 variant, P = 0.00271944) as well as for protein functional impact with ExPASy, Protter, Phyre2, SMART, PROVEAN, SIFT, and I-TASSER, confirming the reported associations. Fam1 associated variant, deemed W31, alters the protein sequence, leading to an early stop codon truncating the normal amino acid sequence from 972 to just 115 amino acids. Fam2 associated variant, deemed W32, may have a subtle impact on receptor function or could be in linkage with a non-coding or regulatory change creating the mild spotting pattern observed in this family.

Metabogenomics reveals four candidate regions involved in the pathophysiology of Equine Metabolic Syndrome.

An analogous condition to human metabolic syndrome, Equine Metabolic Syndrome (EMS) is defined by several clinical signs including obesity, hyperinsulinemia, and peripheral insulin dysregulation (ID). Affected horses may also exhibit hypertension, hyperlipemia and systemic inflammation. Measures of ID typically comprise the gold-standard for diagnosis in veterinary care. Yet, the dynamic nature of insulin homeostasis and complex procedures of typical assays make accurate quantification of ID and EMS challenging. This work aimed to investigate new strategies for identification of biochemical markers and correlated genes in EMS. To quantify EMS risk within this population, we utilized a composite score derived from nine common diagnostic variables. We applied a global liquid chromatography/mass spectroscopy approach (HPLC/MS) to whole plasma collected from 49 Arabian horses, resulting in 3392 high-confidence features and identification of putative metabolites in public databases. We performed a genome wide association analysis with genotypes from the 670k Affymetrix Equine SNP array utilizing EMS-correlated metabolites as phenotypes. We discovered four metabolite features significantly correlated with EMS score (P < 1.474 × 10-5). GWAs for these features results (P = 6.787 × 10-7, Bonferroni) identified four unique candidate regions (r2 > 0.4) containing 63 genes. Significant genomic markers capture 43.52% of the variation in the original EMS score phenotype. The identified genomic loci provide insight into the pathways controlling variation in EMS and the origin of genetic predisposition to the condition. Rapid, feasible and accurate diagnostic tools derived from metabogenomics can be translated into measurable benefits in the timeline and quality of preventative management practices to preserve health in horses.

Genetics of Equine Behavioral Traits.

Behavior is a valuable quantitative trait in the horse because of its impact on performance, work, recreation, and prerequisite close interactions with humans. This article reviews what is known about the genetics of behavior in horses with an emphasis on the genetic basis for temperament traits, neuroendocrine function, and stereotypic behavior. The importance of using modern molecular genetic techniques to the study of equine behavior and recommendations for future research are also discussed. Ultimately, these studies enhance the understanding of the biology of behavior in the horse, improve handler and rider safety, and benefit horse welfare.

Genetics and Signaling Pathways of Laminitis.

Laminitis is a devastating disease with diverse etiologies and few, if any, effective treatments. Gene expression and hypothesis-generating genomic studies have provided a fresh look at the key molecular players at crucial timepoints in diverse experimental and naturally affected tissues. We summarize findings to date, and propose a unifying model of the laminitis disease process that includes several pathogenesis concepts shared with other diseases of epidermal and epithelial tissues. The value of these new pathways as potential therapeutic targets is exciting but will require careful future work to validate new methods and launch systematic clinical trials.

Identification of loci affecting sexually dimorphic patterns for height and recurrent laryngeal neuropathy risk in American Belgian Draft Horses.

Equine recurrent laryngeal neuropathy (RLN) is a bilateral mononeuropathy with an unknown etiology. In Thoroughbreds (TB), we previously demonstrated that the haplotype association for height (LCORL/NCAPG locus on ECA3, which affects body size) and RLN was coincident. In the present study, we performed a genome-wide association scan (GWAS) for RLN in 458 American Belgian Draft Horses, a breed fixed for the LCORL/NCAPG risk alelle. In this breed, RLN risk is associated with sexually dimorphic differences in height, and we identified a novel locus contributing to height in a sex-specific manner: MYPN (ECA1). Yet this specific locus contributes little to RLN risk, suggesting that other growth traits correlated to height may underlie the correlation to this disease. Controlling for height, we identified a locus on ECA15 contributing to RLN risk specifically in males. These results suggest that loci with sex-specific gene expression play an important role in altering growth traits impacting RLN etiology, but not necessarily adult height. These newly identified genes are promising targets for novel preventative and treatment strategies.

Population genomics and morphometric assignment of western honey bees (Apis mellifera L.) in the Republic of South Africa.

BACKGROUNDS: Apis mellifera scutellata and A.m. capensis (the Cape honey bee) are western honey bee subspecies indigenous to the Republic of South Africa (RSA). Both bees are important for biological and economic reasons. First, A.m. scutellata is the invasive "African honey bee" of the Americas and exhibits a number of traits that beekeepers consider undesirable. They swarm excessively, are prone to absconding (vacating the nest entirely), usurp other honey bee colonies, and exhibit heightened defensiveness. Second, Cape honey bees are socially parasitic bees; the workers can reproduce thelytokously. Both bees are indistinguishable visually. Therefore, we employed Genotyping-by-Sequencing (GBS), wing geometry and standard morphometric approaches to assess the genetic diversity and population structure of these bees to search for diagnostic markers that can be employed to distinguish between the two subspecies. RESULTS: Apis mellifera scutellata possessed the highest mean number of polymorphic SNPs (among 2449 informative SNPs) with minor allele frequencies > 0.05 (Np = 88%). The RSA honey bees generated a high level of expected heterozygosity (Hexp = 0.24). The mean genetic differentiation (FST; 6.5%) among the RSA honey bees revealed that approximately 93% of the genetic variation was accounted for within individuals of these subspecies. Two genetically distinct clusters (K = 2) corresponding to both subspecies were detected by Model-based Bayesian clustering and supported by Principal Coordinates Analysis (PCoA) inferences. Selected highly divergent loci (n = 83) further reinforced a distinctive clustering of two subspecies across geographical origins, accounting for approximately 83% of the total variation in the PCoA plot. The significant correlation of allele frequencies at divergent loci with environmental variables suggested that these populations are adapted to local conditions. Only 17 of 48 wing geometry and standard morphometric parameters were useful for clustering A.m. capensis, A.m. scutellata, and hybrid individuals. CONCLUSIONS: We produced a minimal set of 83 SNP loci and 17 wing geometry and standard morphometric parameters useful for identifying the two RSA honey bee subspecies by genotype and phenotype. We found that genes involved in neurology/behavior and development/growth are the most prominent heritable traits evolved in the functional evolution of honey bee populations in RSA. These findings provide a starting point for understanding the functional basis of morphological differentiations and ecological adaptations of the two honey bee subspecies in RSA.

Developing a 670k genotyping array to tag ~2M SNPs across 24 horse breeds.

BACKGROUND: To date, genome-scale analyses in the domestic horse have been limited by suboptimal single nucleotide polymorphism (SNP) density and uneven genomic coverage of the current SNP genotyping arrays. The recent availability of whole genome sequences has created the opportunity to develop a next generation, high-density equine SNP array. RESULTS: Using whole genome sequence from 153 individuals representing 24 distinct breeds collated by the equine genomics community, we cataloged over 23 million de novo discovered genetic variants. Leveraging genotype data from individuals with both whole genome sequence, and genotypes from lower-density, legacy SNP arrays, a subset of ~5 million high-quality, high-density array candidate SNPs were selected based on breed representation and uniform spacing across the genome. Considering probe design recommendations from a commercial vendor (Affymetrix, now Thermo Fisher Scientific) a set of ~2 million SNPs were selected for a next-generation high-density SNP chip (MNEc2M). Genotype data were generated using the MNEc2M array from a cohort of 332 horses from 20 breeds and a lower-density array, consisting of ~670 thousand SNPs (MNEc670k), was designed for genotype imputation. CONCLUSIONS: Here, we document the steps taken to design both the MNEc2M and MNEc670k arrays, report genomic and technical properties of these genotyping platforms, and demonstrate the imputation capabilities of these tools for the domestic horse.

Polymorphism at expressed DQ and DR loci in five common equine MHC haplotypes.

The polymorphism of major histocompatibility complex (MHC) class II DQ and DR genes in five common equine leukocyte antigen (ELA) haplotypes was determined through sequencing of mRNA transcripts isolated from lymphocytes of eight ELA homozygous horses. Ten expressed MHC class II genes were detected in horses of the ELA-A3 haplotype carried by the donor horses of the equine bacterial artificial chromosome (BAC) library and the reference genome sequence: four DR genes and six DQ genes. The other four ELA haplotypes contained at least eight expressed polymorphic MHC class II loci. Next generation sequencing (NGS) of genomic DNA of these four MHC haplotypes revealed stop codons in the DQA3 gene in the ELA-A2, ELA-A5, and ELA-A9 haplotypes. Few NGS reads were obtained for the other MHC class II genes that were not amplified in these horses. The amino acid sequences across haplotypes contained locus-specific residues, and the locus clusters produced by phylogenetic analysis were well supported. The MHC class II alleles within the five tested haplotypes were largely non-overlapping between haplotypes. The complement of equine MHC class II DQ and DR genes appears to be well conserved between haplotypes, in contrast to the recently described variation in class I gene loci between equine MHC haplotypes. The identification of allelic series of equine MHC class II loci will aid comparative studies of mammalian MHC conservation and evolution and may also help to interpret associations between the equine MHC class II region and diseases of the horse.

Application of Genomic Estimation Methods of Inbreeding and Population Structure in an Arabian Horse Herd.

Horse breeders rely heavily on pedigrees for identification of ancestry in breeding stock. Inaccurate pedigrees may erroneously assign individuals to false lineages or breed memberships resulting in wrong estimates of inbreeding and coancestry. Moreover, discrepancies in pedigree records can lead breeders seeking to limit inbreeding into making misguided breeding decisions. Genome-wide SNPs provide a quantitative tool to aid in the resolution of lineage assignments and the calculation of genomic measures of relatedness. The aim of this project was to pilot a comparison between pedigree and genomic relatedness and inbreeding measures in a herd of 36 pedigreed Egyptian Arabian horses genotyped using the Equine SNP70 platform (Geneseek, Inc.). Moreover, we sought to estimate the minimum number of markers sufficient for genomic inbreeding calculations. Pedigree inbreeding values were moderately correlated with genomic inbreeding values (r = 0.406), whereas genomic relationships and pedigree relationships have a high correlation (r = 0.77). Although first degree relationships were successfully reconstructed, more distant relationships were difficult to resolve. Multi-dimensional scaling and clustering analysis agreed with within-herd pedigree information. In comparing the herd to a reference sample of United States, Polish, and Egyptian Arabian horses, the herd's historically recorded Egyptian lineage was successfully recovered. We conclude that genomic estimates of inbreeding and relationships are superior to their pedigree counterparts. They can be thus utilized in conservation of valuable lines of livestock, and in breeds at risk for loss of genomic diversity. We also postulate a minimum of 2000 markers in linkage equilibrium to be used for inbreeding estimation.

Host genetic influence on papillomavirus-induced tumors in the horse.

The common equine skin tumors known as sarcoids have been causally associated with infection by bovine papillomavirus (BPV). Additionally, there is evidence for host genetic susceptibility to sarcoids. We investigated the genetic basis of susceptibility to sarcoid tumors on a cohort of 82 affected horses and 270 controls genotyped on a genome-wide platform and two custom panels. A Genome Wide Association Study (GWAS) identified candidate regions on six chromosomes. Bayesian probability analysis of the same dataset verified only the regions on equine chromosomes (ECA) 20 and 22. Fine mapping using custom-produced SNP arrays for ECA20 and ECA22 regions identified two marker loci with high levels of significance: SNP BIEC2-530826 (map position 32,787,619) on ECA20 in an intron of the DQA1 gene in the Major Histocompatibility Complex (MHC) class II region (p = 4.6e-06), and SNP BIEC2-589604 (map position 25,951,536) on ECA22 in a 200 kb region containing four candidate genes: PROCR, EDEM2, EIF6 and MMP24 (p = 2.14e-06). The marker loci yielded odds ratios of 5.05 and 4.02 for ECA20 and ECA22, respectively. Associations between genetic MHC class II variants and papillomavirus-induced tumors have been reported for human papillomavirus and cottontail rabbit papillomavirus infections. This suggests a common mechanism for susceptibility to tumor progression that may involve subversion of the host immune response. This study also identified a genomic region other than MHC that influenced papillomavirus-induced tumor development in the studied population.

The MC1R and ASIP Coat Color Loci May Impact Behavior in the Horse.

Shared signaling pathways utilized by melanocytes and neurons result in pleiotropic traits of coat color and behavior in many mammalian species. For example, in humans polymorphisms at MC1R cause red hair, increased heat sensitivity, and lower pain tolerance. In deer mice, rats, and foxes, ASIP polymorphisms causing black coat color lead to more docile demeanors and reduced activity. Horse (Equus caballus) base coat color is primarily determined by polymorphisms at the Melanocortin-1 Receptor (MC1R) and Agouti Signaling Protein (ASIP) loci, creating a black, bay, or chestnut coat. Our goal was to investigate correlations between genetic loci for coat color and temperament traits in the horse. We genotyped a total of 215 North American Tennessee Walking Horses for the 2 most common alleles at the MC1R (E/e) and ASIP (A/a) loci using previously published PCR and RFLP methods. The horses had a mean age of 10.5 years and comprised 83 geldings, 25 stallions, and 107 mares. To assess behavior, we adapted a previously published survey for handlers to score horses from 1 to 9 on 20 questions related to specific aspects of temperament. We utilized principle component analysis to combine the individual survey scores into 4 factors of variation in temperament phenotype. A factor component detailing self-reliance correlated with genotypes at the ASIP locus; black mares (aa) were more independent than bay mares (A_) (P = 0.0063). These findings illuminate a promising and novel animal model for future study of neuroendocrine mechanisms in complex behavioral phenotypes.

Generation of a de novo transcriptome from equine lamellar tissue.

BACKGROUND: Laminitis, the structural failure of interdigitated tissue that suspends the distal skeleton within the hoof capsule, is a devastating disease that is the second leading cause of both lameness and euthanasia in the horse. Current transcriptomic research focuses on the expression of known genes. However, as this tissue is quite unique and equine gene annotation is largely derived from computational predictions, there are likely yet uncharacterized transcripts that may be involved in the etiology of laminitis. In order to create a novel annotation resource, we performed whole transcriptome sequencing of sagittal lamellar sections from one control and two laminitis affected horses. RESULTS: Whole transcriptome sequencing of the three samples resulted in 113 million reads. Overall, 88 % of the reads mapped to the equCab2 reference genome, allowing for the identification of 119,430 SNPs. The de novo assembly generated around 75,000 transcripts, of which 36,000 corresponded to known annotations. Annotated transcript models are hosted in a public data repository and thus can be easily accessed or loaded into genome browsers. RT-PCR of 12 selected assemblies confirmed structure and expression in lamellar tissue. CONCLUSIONS: Transcriptome sequencing represents a powerful tool to expand on equine annotation and identify novel targets for further laminitis research.

Mutations in MITF and PAX3 cause "splashed white" and other white spotting phenotypes in horses.

During fetal development neural-crest-derived melanoblasts migrate across the entire body surface and differentiate into melanocytes, the pigment-producing cells. Alterations in this precisely regulated process can lead to white spotting patterns. White spotting patterns in horses are a complex trait with a large phenotypic variance ranging from minimal white markings up to completely white horses. The "splashed white" pattern is primarily characterized by an extremely large blaze, often accompanied by extended white markings at the distal limbs and blue eyes. Some, but not all, splashed white horses are deaf. We analyzed a Quarter Horse family segregating for the splashed white coat color. Genome-wide linkage analysis in 31 horses gave a positive LOD score of 1.6 in a region on chromosome 6 containing the PAX3 gene. However, the linkage data were not in agreement with a monogenic inheritance of a single fully penetrant mutation. We sequenced the PAX3 gene and identified a missense mutation in some, but not all, splashed white Quarter Horses. Genome-wide association analysis indicated a potential second signal near MITF. We therefore sequenced the MITF gene and found a 10 bp insertion in the melanocyte-specific promoter. The MITF promoter variant was present in some splashed white Quarter Horses from the studied family, but also in splashed white horses from other horse breeds. Finally, we identified two additional non-synonymous mutations in the MITF gene in unrelated horses with white spotting phenotypes. Thus, several independent mutations in MITF and PAX3 together with known variants in the EDNRB and KIT genes explain a large proportion of horses with the more extreme white spotting phenotypes.

Genomic analysis establishes correlation between growth and laryngeal neuropathy in Thoroughbreds.

BACKGROUND: Equine recurrent laryngeal neuropathy (RLN) is a bilateral mononeuropathy with an unknown pathogenesis that significantly affects performance in Thoroughbreds. A genetic contribution to the pathogenesis of RLN is suggested by the higher prevalence of the condition in offspring of RLN-affected than unaffected stallions. To better understand RLN pathogenesis and its genetic basis, we performed a genome-wide association (GWAS) of 282 RLN-affected and 268 control Thoroughbreds. RESULTS: We found a significant association of RLN with the LCORL/NCAPG locus on ECA3 previously shown to affect body size in horses. Using height at the withers of 505 of these horses, we confirmed the strong association of this locus with body size, and demonstrated a significant phenotypic and genetic correlation between height and RLN grade in this cohort. Secondary genetic associations for RLN on ECA18 and X did not correlate with withers height in our cohort, but did contain candidate genes likely influencing muscle physiology and growth: myostatin (MSTN) and integral membrane protein 2A (ITM2A). CONCLUSIONS: This linkage between body size and RLN suggests that selective breeding to reduce RLN prevalence would likely reduce adult size in this population. However, our results do not preclude the possibility of modifier loci that attenuate RLN risk without reducing size or performance, or that the RLN risk allele is distinct but tightly linked to the body size locus on ECA3. This study is both the largest body size GWAS and the largest RLN GWAS within Thoroughbred horses to date, and suggests that improved understanding of the relationship between genetics, equine growth rate, and RLN prevalence may significantly advance our understanding and management of this disease.