Dog10K_Boxer_Tasha_1.0: A Long-Read Assembly of the Dog Reference Genome

Large-scale genomic analysis of the domestic dog informs biological discovery

Recent advances in genomics, coupled with a unique population structure and remarkable levels of variation, have propelled the domestic dog to new levels as a system for understanding fundamental principles in mammalian biology. Central to this advance are more than 350 recognized breeds, each a closed population that has undergone selection for unique features. Genetic variation in the domestic dog is particularly well characterized compared with other domestic mammals, with almost 3000 high-coverage genomes publicly available. Importantly, as the number of sequenced genomes increases, new avenues for analysis are becoming available. Herein, we discuss recent discoveries in canine genomics regarding behavior, morphology, and disease susceptibility. We explore the limitations of current data sets for variant interpretation, tradeoffs between sequencing strategies, and the burgeoning role of long-read genomes for capturing structural variants. In addition, we consider how large-scale collections of whole-genome sequence data drive rare variant discovery and assess the geographic distribution of canine diversity, which identifies Asia as a major source of missing variation. Finally, we review recent comparative genomic analyses that will facilitate annotation of the noncoding genome in dogs.

Duplications and Retrogenes Are Numerous and Widespread in Modern Canine Genomic Assemblies

Recent years have seen a dramatic increase in the number of canine genome assemblies available. Duplications are an important source of evolutionary novelty and are also prone to misassembly. We explored the duplication content of nine canine genome assemblies using both genome self-alignment and read-depth approaches. We find that 8.58% of the genome is duplicated in the canFam4 assembly, derived from the German Shepherd Dog Mischka, including 90.15% of unplaced contigs. Highlighting the continued difficulty in properly assembling duplications, less than half of read-depth and assembly alignment duplications overlap, but the mCanLor1.2 Greenland wolf assembly shows greater concordance. Further study shows the presence of multiple segments that have alignments to four or more duplicate copies. These high-recurrence duplications correspond to gene retrocopies. We identified 3,892 candidate retrocopies from 1,316 parental genes in the canFam4 assembly and find that ∼8.82% of duplicated base pairs involve a retrocopy, confirming this mechanism as a major driver of gene duplication in canines. Similar patterns are found across eight other recent canine genome assemblies, with metrics supporting a greater quality of the PacBio HiFi mCanLor1.2 assembly. Comparison between the wolf and other canine assemblies found that 92% of retrocopy insertions are shared between assemblies. By calculating the number of generations since genome divergence, we estimate that new retrocopy insertions appear, on average, in 1 out of 3,514 births. Our analyses illustrate the impact of retrogene formation on canine genomes and highlight the variable representation of duplicated sequences among recently completed canine assemblies.

Early Injury Landscape in Vein Harvest by Single-Cell and Spatial Transcriptomics

BACKGROUND

Vein graft failure following cardiovascular bypass surgery results in significant patient morbidity and cost to the healthcare system. Vein graft injury can occur during autogenous vein harvest and preparation, as well as after implantation into the arterial system, leading to the development of intimal hyperplasia, vein graft stenosis, and, ultimately, bypass graft failure. Although previous studies have identified maladaptive pathways that occur shortly after implantation, the specific signaling pathways that occur during vein graft preparation are not well defined and may result in a cumulative impact on vein graft failure. We, therefore, aimed to elucidate the response of the vein conduit wall during harvest and following implantation, probing the key maladaptive pathways driving graft failure with the overarching goal of identifying therapeutic targets for biologic intervention to minimize these natural responses to surgical vein graft injury.

METHODS

Employing a novel approach to investigating vascular pathologies, we harnessed both single-nuclei RNA-sequencing and spatial transcriptomics analyses to profile the genomic effects of vein grafts after harvest and distension, then compared these findings to vein grafts obtained 24 hours after carotid-carotid vein bypass implantation in a canine model (n=4).

RESULTS

Spatial transcriptomic analysis of canine cephalic vein after initial conduit harvest and distention revealed significant enrichment of pathways (P<0.05) involved in the activation of endothelial cells (ECs), fibroblasts, and vascular smooth muscle cells, namely pathways responsible for cellular proliferation and migration and platelet activation across the intimal and medial layers, cytokine signaling within the adventitial layer, and ECM (extracellular matrix) remodeling throughout the vein wall. Subsequent single-nuclei RNA-sequencing analysis supported these findings and further unveiled distinct EC and fibroblast subpopulations with significant upregulation (P<0.05) of markers related to endothelial injury response and cellular activation of ECs, fibroblasts, and vascular smooth muscle cells. Similarly, in vein grafts obtained 24 hours after arterial bypass, there was an increase in myeloid cell, protomyofibroblast, injury response EC, and mesenchymal-transitioning EC subpopulations with a concomitant decrease in homeostatic ECs and fibroblasts. Among these markers were genes previously implicated in vein graft injury, including VCAN, FBN1, and VEGFC, in addition to novel genes of interest, such as GLIS3 and EPHA3. These genes were further noted to be driving the expression of genes implicated in vascular remodeling and graft failure, such as IL-6, TGFBR1, SMAD4, and ADAMTS9. By integrating the spatial transcriptomics and single-nuclei RNA-sequencing data sets, we highlighted the spatial architecture of the vein graft following distension, wherein activated and mesenchymal-transitioning ECs, myeloid cells, and fibroblasts were notably enriched in the intima and media of distended veins. Finally, intercellular communication network analysis unveiled the critical roles of activated ECs, mesenchymal-transitioning ECs, protomyofibroblasts, and vascular smooth muscle cells in upregulating signaling pathways associated with cellular proliferation (MDK [midkine], PDGF [platelet-derived growth factor], VEGF [vascular endothelial growth factor]), transdifferentiation (Notch), migration (ephrin, semaphorin), ECM remodeling (collagen, laminin, fibronectin), and inflammation (thrombospondin), following distension.

CONCLUSIONS

Vein conduit harvest and distension elicit a prompt genomic response facilitated by distinct cellular subpopulations heterogeneously distributed throughout the vein wall. This response was found to be further exacerbated following vein graft implantation, resulting in a cascade of maladaptive gene regulatory networks. Together, these results suggest that distension initiates the upregulation of pathological pathways that may ultimately contribute to bypass graft failure and presents potential early targets warranting investigation for targeted therapies. This work highlights the first applications of single-nuclei and spatial transcriptomic analyses to investigate venous pathologies, underscoring the utility of these methodologies and providing a foundation for future investigations.

Integrated single-nuclei and spatial transcriptomic analysis reveals propagation of early acute vein harvest and distension injury signaling pathways following arterial implantation

Background

Vein graft failure (VGF) following cardiovascular bypass surgery results in significant patient morbidity and cost to the healthcare system. Vein graft injury can occur during autogenous vein harvest and preparation, as well as after implantation into the arterial system, leading to the development of intimal hyperplasia, vein graft stenosis, and, ultimately, bypass graft failure. While previous studies have identified maladaptive pathways that occur shortly after implantation, the specific signaling pathways that occur during vein graft preparation are not well defined and may result in a cumulative impact on VGF. We, therefore, aimed to elucidate the response of the vein conduit wall during harvest and following implantation, probing the key maladaptive pathways driving graft failure with the overarching goal of identifying therapeutic targets for biologic intervention to minimize these natural responses to surgical vein graft injury.

Methods

Employing a novel approach to investigating vascular pathologies, we harnessed both single-nuclei RNA-sequencing (snRNA-seq) and spatial transcriptomics (ST) analyses to profile the genomic effects of vein grafts after harvest and distension, then compared these findings to vein grafts obtained 24 hours after carotid-cartoid vein bypass implantation in a canine model (n=4).

Results

Spatial transcriptomic analysis of canine cephalic vein after initial conduit harvest and distention revealed significant enrichment of pathways (P < 0.05) involved in the activation of endothelial cells (ECs), fibroblasts (FBs), and vascular smooth muscle cells (VSMCs), namely pathways responsible for cellular proliferation and migration and platelet activation across the intimal and medial layers, cytokine signaling within the adventitial layer, and extracellular matrix (ECM) remodeling throughout the vein wall. Subsequent snRNA-seq analysis supported these findings and further unveiled distinct EC and FB subpopulations with significant upregulation (P < 0.00001) of markers related to endothelial injury response and cellular activation of ECs, FBs, and VSMCs. Similarly, in vein grafts obtained 24 hours after arterial bypass, there was an increase in myeloid cell, protomyofibroblast, injury-response EC, and mesenchymal-transitioning EC subpopulations with a concomitant decrease in homeostatic ECs and fibroblasts. Among these markers were genes previously implicated in vein graft injury, including VCAN (versican), FBN1 (fibrillin-1), and VEGFC (vascular endothelial growth factor C), in addition to novel genes of interest such as GLIS3 (GLIS family zinc finger 3) and EPHA3 (ephrin-A3). These genes were further noted to be driving the expression of genes implicated in vascular remodeling and graft failure, such as IL-6, TGFBR1, SMAD4, and ADAMTS9. By integrating the ST and snRNA-seq datasets, we highlighted the spatial architecture of the vein graft following distension, wherein activated and mesenchymal-transitioning ECs, myeloid cells, and FBs were notably enriched in the intima and media of distended veins. Lastly, intercellular communication network analysis unveiled the critical roles of activated ECs, mesenchymal transitioning ECs, protomyofibroblasts, and VSMCs in upregulating signaling pathways associated with cellular proliferation (MDK, PDGF, VEGF), transdifferentiation (Notch), migration (ephrin, semaphorin), ECM remodeling (collagen, laminin, fibronectin), and inflammation (thrombospondin), following distension.

Conclusions

Vein conduit harvest and distension elicit a prompt genomic response facilitated by distinct cellular subpopulations heterogeneously distributed throughout the vein wall. This response was found to be further exacerbated following vein graft implantation, resulting in a cascade of maladaptive gene regulatory networks. Together, these results suggest that distension initiates the upregulation of pathological pathways that may ultimately contribute to bypass graft failure and presents potential early targets warranting investigation for targeted therapies. This work highlights the first applications of single-nuclei and spatial transcriptomic analyses to investigate venous pathologies, underscoring the utility of these methodologies and providing a foundation for future investigations.

Genome-wide methylation patterns from canine nanopore assemblies

Recent advances in long-read sequencing have enabled the creation of reference-quality genome assemblies for multiple individuals within a species. In particular, 8 long-read genome assemblies have recently been published for the canine model (dogs and wolves). These assemblies were created using a range of sequencing and computational approaches, with only limited comparisons described among subsets of the assemblies. Here we present 3 high-quality de novo reference assemblies based upon Oxford Nanopore long-read sequencing: 2 Bernese Mountain Dogs (BD & OD) and a Cairn terrier (CA611). These breeds are of particular interest due to the enrichment of unresolved genetic disorders. Leveraging advancement in software technologies, we utilized published data of Labrador Retriever (Yella) to generate a new assembly, resulting in a ∼280-fold increase in continuity (N50 size of 91 kbp vs 25.75 Mbp). In conjunction with these 4 new assemblies, we uniformly assessed 8 existing assemblies for generalized quality metrics, sequence divergence, and a detailed BUSCO assessment. We identified a set of ∼400 conserved genes during the BUSCO analysis missing in all assemblies. Genome-wide methylation profiles were generated from the nanopore sequencing, resulting in broad concordance with existing whole-genome and reduced-representation bisulfite sequencing, while highlighting superior overage of mobile elements. These analyses demonstrate the ability of Nanopore sequencing to resolve the sequence and epigenetic profile of canine genomes.

Standing in the canine precision medicine knowledge gap: Improving annotation of canine cancer genomic biomarkers through systematic comparative analysis of human cancer mutations in COSMIC

The accrual of cancer mutation data and related functional and clinical associations have revolutionised human oncology, enabling the advancement of precision medicine and biomarker-guided clinical management. The catalogue of cancer mutations is also growing in canine cancers. However, without direct high-powered functional data in dogs, it remains challenging to interpret and utilise them in research and clinical settings. It is well-recognised that canine and human cancers share genetic, molecular and phenotypic similarities. Therefore, leveraging the massive wealth of human mutation data may help advance canine oncology. Here, we present a structured analysis of sequence conservation and conversion of human mutations to the canine genome through a 'caninisation' process. We applied this analysis to COSMIC, the Catalogue of Somatic Mutations in Cancer, the most prominent human cancer mutation database. For the project's initial phase, we focused on the subset of the COSMIC data corresponding to Cancer Gene Census (CGC) genes. A total of 670 canine orthologs were found for 721 CGC genes. In these genes, 365 K unique mutations across 160 tumour types were converted successfully to canine coordinates. We identified shared putative cancer-driving mutations, including pathogenic and hotspot mutations and mutations bearing similar biomarker associations with diagnostic, prognostic and therapeutic utility. Thus, this structured caninisation of human cancer mutations facilitates the interpretation and annotation of canine mutations and helps bridge the knowledge gap to enable canine precision medicine.

Genome sequencing of 2000 canids by the Dog10K consortium advances the understanding of demography, genome function and architecture

BACKGROUND

The international Dog10K project aims to sequence and analyze several thousand canine genomes. Incorporating 20 × data from 1987 individuals, including 1611 dogs (321 breeds), 309 village dogs, 63 wolves, and four coyotes, we identify genomic variation across the canid family, setting the stage for detailed studies of domestication, behavior, morphology, disease susceptibility, and genome architecture and function.

RESULTS

We report the analysis of > 48 M single-nucleotide, indel, and structural variants spanning the autosomes, X chromosome, and mitochondria. We discover more than 75% of variation for 239 sampled breeds. Allele sharing analysis indicates that 94.9% of breeds form monophyletic clusters and 25 major clades. German Shepherd Dogs and related breeds show the highest allele sharing with independent breeds from multiple clades. On average, each breed dog differs from the UU_Cfam_GSD_1.0 reference at 26,960 deletions and 14,034 insertions greater than 50 bp, with wolves having 14% more variants. Discovered variants include retrogene insertions from 926 parent genes. To aid functional prioritization, single-nucleotide variants were annotated with SnpEff and Zoonomia phyloP constraint scores. Constrained positions were negatively correlated with allele frequency. Finally, the utility of the Dog10K data as an imputation reference panel is assessed, generating high-confidence calls across varied genotyping platform densities including for breeds not included in the Dog10K collection.

CONCLUSIONS

We have developed a dense dataset of 1987 sequenced canids that reveals patterns of allele sharing, identifies likely functional variants, informs breed structure, and enables accurate imputation. Dog10K data are publicly available.

Common Orthopedic Traits and Screening for Breeding Programs

Orthopedic diseases are complex traits, meaning genetics and environmental factors affect risk, making identification of genetic associations difficult. In the United States, hip and elbow scores, patellar luxation scores, Legg-Calvé-Perthes disease, and shoulder osteochondrosis affectedness are available in the Orthopedic Foundation for Animals registry. Distraction indices and extended, ventrodorsal hip conformation scores are recorded by PennHIP. Application of estimated breeding values for hip and elbow dysplasia in breeder selection reduces the severity and prevalence of these traits. Genomic prediction and whole-genome sequence technologies and methods should improve knowledge of genetics underlying orthopedic diseases, leading to improved canine orthopedic genetic quality.

The Australasian dingo archetype: de novo chromosome-length genome assembly, DNA methylome, and cranial morphology

BACKGROUND

One difficulty in testing the hypothesis that the Australasian dingo is a functional intermediate between wild wolves and domesticated breed dogs is that there is no reference specimen. Here we link a high-quality de novo long-read chromosomal assembly with epigenetic footprints and morphology to describe the Alpine dingo female named Cooinda. It was critical to establish an Alpine dingo reference because this ecotype occurs throughout coastal eastern Australia where the first drawings and descriptions were completed.

FINDINGS

We generated a high-quality chromosome-level reference genome assembly (Canfam_ADS) using a combination of Pacific Bioscience, Oxford Nanopore, 10X Genomics, Bionano, and Hi-C technologies. Compared to the previously published Desert dingo assembly, there are large structural rearrangements on chromosomes 11, 16, 25, and 26. Phylogenetic analyses of chromosomal data from Cooinda the Alpine dingo and 9 previously published de novo canine assemblies show dingoes are monophyletic and basal to domestic dogs. Network analyses show that the mitochondrial DNA genome clusters within the southeastern lineage, as expected for an Alpine dingo. Comparison of regulatory regions identified 2 differentially methylated regions within glucagon receptor GCGR and histone deacetylase HDAC4 genes that are unmethylated in the Alpine dingo genome but hypermethylated in the Desert dingo. Morphologic data, comprising geometric morphometric assessment of cranial morphology, place dingo Cooinda within population-level variation for Alpine dingoes. Magnetic resonance imaging of brain tissue shows she had a larger cranial capacity than a similar-sized domestic dog.

CONCLUSIONS

These combined data support the hypothesis that the dingo Cooinda fits the spectrum of genetic and morphologic characteristics typical of the Alpine ecotype. We propose that she be considered the archetype specimen for future research investigating the evolutionary history, morphology, physiology, and ecology of dingoes. The female has been taxidermically prepared and is now at the Australian Museum, Sydney.

Identification of Genetic Risk Factors for Monogenic and Complex Canine Diseases

Advances in DNA sequencing and other technologies have greatly facilitated the identification of genetic risk factors for inherited diseases in dogs. We review recent technological developments based on selected examples from canine disease genetics. The identification of disease-causing variants in dogs with monogenic diseases may become a widely employed diagnostic approach in clinical veterinary medicine in the not-too-distant future. Diseases with complex modes of inheritance continue to pose challenges to researchers but have also become much more tangible than in the past. In addition to strategies for identifying genetic risk factors, we provide some thoughts on the interpretation of sequence variants that are largely inspired by developments in human clinical genetics.

The Australasian dingo archetype: De novo chromosome-length genome assembly, DNA methylome, and cranial morphology

Background

One difficulty in testing the hypothesis that the Australasian dingo is a functional intermediate between wild wolves and domesticated breed dogs is that there is no reference specimen. Here we link a high-quality de novo long read chromosomal assembly with epigenetic footprints and morphology to describe the Alpine dingo female named Cooinda. It was critical to establish an Alpine dingo reference because this ecotype occurs throughout coastal eastern Australia where the first drawings and descriptions were completed.

Findings

We generated a high-quality chromosome-level reference genome assembly (Canfam_ADS) using a combination of Pacific Bioscience, Oxford Nanopore, 10X Genomics, Bionano, and Hi-C technologies. Compared to the previously published Desert dingo assembly, there are large structural rearrangements on Chromosomes 11, 16, 25 and 26. Phylogenetic analyses of chromosomal data from Cooinda the Alpine dingo and nine previously published de novo canine assemblies show dingoes are monophyletic and basal to domestic dogs. Network analyses show that the mtDNA genome clusters within the southeastern lineage, as expected for an Alpine dingo. Comparison of regulatory regions identified two differentially methylated regions within glucagon receptor GCGR and histone deacetylase HDAC4 genes that are unmethylated in the Alpine dingo genome but hypermethylated in the Desert dingo. Morphological data, comprising geometric morphometric assessment of cranial morphology place dingo Cooinda within population-level variation for Alpine dingoes. Magnetic resonance imaging of brain tissue show she had a larger cranial capacity than a similar-sized domestic dog.

Conclusions

These combined data support the hypothesis that the dingo Cooinda fits the spectrum of genetic and morphological characteristics typical of the Alpine ecotype. We propose that she be considered the archetype specimen for future research investigating the evolutionary history, morphology, physiology, and ecology of dingoes. The female has been taxidermically prepared and is now at the Australian Museum, Sydney.

Domestic dog lineages reveal genetic drivers of behavioral diversification

Selective breeding of domestic dogs has generated diverse breeds often optimized for performing specialized tasks. Despite the heritability of breed-typical behavioral traits, identification of causal loci has proven challenging due to the complexity of canine population structure. We overcome longstanding difficulties in identifying genetic drivers of canine behavior by developing a framework for understanding relationships between breeds and the behaviors that define them, utilizing genetic data for over 4,000 domestic, semi-feral, and wild canids and behavioral survey data for over 46,000 dogs. We identify ten major canine genetic lineages and their behavioral correlates and show that breed diversification is predominantly driven by non-coding regulatory variation. We determine that lineage-associated genes converge in neurodevelopmental co-expression networks, identifying a sheepdog-associated enrichment for interrelated axon guidance functions. This work presents a scaffold for canine diversification that positions the domestic dog as an unparalleled system for revealing the genetic origins of behavioral diversity.

Canine reference genome accuracy impacts variant calling: Lessons learned from investigating embryonic lethal variants

Deficient homozygosity of a variant maintained in a population suggests that the variant may be embryonic lethal. We examined whole genome sequence data from 675 canids to investigate for variants with missing homozygosity and high predicted impact. Our analysis identified 45 variants, in 32 genes. However, further scrutiny of the sequence reads revealed that all but one of these variants were artifacts of the variant calling process when using CanFam3.1, a widely utilized canine reference genome. We demonstrate that the use of multiple, newer reference genomes could reduce artifacts and lead to more accurate variant identification.

GENESPACE tracks regions of interest and gene copy number variation across multiple genomes

The development of multiple chromosome-scale reference genome sequences in many taxonomic groups has yielded a high-resolution view of the patterns and processes of molecular evolution. Nonetheless, leveraging information across multiple genomes remains a significant challenge in nearly all eukaryotic systems. These challenges range from studying the evolution of chromosome structure, to finding candidate genes for quantitative trait loci, to testing hypotheses about speciation and adaptation. Here, we present GENESPACE, which addresses these challenges by integrating conserved gene order and orthology to define the expected physical position of all genes across multiple genomes. We demonstrate this utility by dissecting presence–absence, copy-number, and structural variation at three levels of biological organization: spanning 300 million years of vertebrate sex chromosome evolution, across the diversity of the Poaceae (grass) plant family, and among 26 maize cultivars. The methods to build and visualize syntenic orthology in the GENESPACE R package offer a significant addition to existing gene family and synteny programs, especially in polyploid, outbred, and other complex genomes.

The genome is the complete DNA sequence of an individual. It is a crucial foundation for many studies in medicine, agriculture, and conservation biology. Advances in genetics have made it possible to rapidly sequence, or read out, the genome of many organisms. For closely related species, scientists can then do detailed comparisons, revealing similar genes with a shared past or a common role, but comparing more distantly related organisms remains difficult.

One major challenge is that genes are often lost or duplicated over evolutionary time. One way to be more confident is to look at ‘synteny’, or how genes are organized or ordered within the genome. In some groups of species, synteny persists across millions of years of evolution. Combining sequence similarity with gene order could make comparisons between distantly related species more robust.

To do this, Lovell et al. developed GENESPACE, a software that links similarities between DNA sequences to the order of genes in a genome. This allows researchers to visualize and explore related DNA sequences and determine whether genes have been lost or duplicated. To demonstrate the value of GENESPACE, Lovell et al. explored evolution in vertebrates and flowering plants. The software was able to highlight the shared sequences between unique sex chromosomes in birds and mammals, and it was able to track the positions of genes important in the evolution of grass crops including maize, wheat, and rice.

Exploring the genetic code in this way could lead to a better understanding of the evolution of important sections of the genome. It might also allow scientists to find target genes for applications like crop improvement. Lovell et al. have designed the GENESPACE software to be easy for other scientists to use, allowing them to make graphics and perform analyses with few programming skills.

Dog-human translational genomics: state of the art and genomic resources

Innovative models for medical research are strongly required nowadays. Convincing evidence supports dog as the most suitable spontaneous model for several human genetic diseases. Decades of studies on dog genome allowed the identification of hundreds of mutations causing genetic disorders, many of which are proposed as counterparts responsible for human diseases. Traditionally, the murine model is the most extensively used in human translational research. However, this species shows large physiological differences from humans, and it is kept under a controlled artificial environment. Conversely, canine genetic disorders often show pathophysiological and clinical features highly resembling the human counterpart. In addition, dogs share the same environment with humans; therefore, they are naturally exposed to many risk factors. Thus, different branches of translational medicine aim to study spontaneously occurring diseases in dogs to provide a more reliable model for human disorders. This review offers a comprehensive overview of the knowledge and resources available today for all the researchers involved in the field of dog-human translational medicine. Some of the main successful examples from dog-human translational genomics are reported, such as the canine association studies which helped to identify the causal mutation in the human counterpart. We also illustrated the ongoing projects aiming to create public canine big datasets. Finally, specific online databases are discussed along with several information resources that can speed up clinical translational research.

Cross-species enhancer prediction using machine learning

Cis-regulatory elements (CREs) are non-coding parts of the genome that play a critical role in gene expression regulation. Enhancers, as an important example of CREs, interact with genes to influence complex traits like disease, heat tolerance and growth rate. Much of what is known about enhancers come from studies of humans and a few model organisms like mouse, with little known about other mammalian species. Previous studies have attempted to identify enhancers in less studied mammals using comparative genomics but with limited success. Recently, Machine Learning (ML) techniques have shown promising results to predict enhancer regions. Here, we investigated the ability of ML methods to identify enhancers in three non-model mammalian species (cattle, pig and dog) using human and mouse enhancer data from VISTA and publicly available ChIP-seq. We tested nine models, using four different representations of the DNA sequences in cross-species prediction using both the VISTA dataset and species-specific ChIP-seq data. We identified between 809,399 and 877,278 enhancer-like regions (ELRs) in the study species (11.6-13.7% of each genome). These predictions were close to the ~8% proportion of ELRs that covered the human genome. We propose that our ML methods have predictive ability for identifying enhancers in non-model mammalian species. We have provided a list of high confidence enhancers at https://github.com/DaviesCentreInformatics/Cross-species-enhancer-prediction and believe these enhancers will be of great use to the community.