Ancient genomes reveal over two thousand years of dingo population structure

Dingoes (Canis familiaris dingo), are significant ecologically and culturally in Australia. They are the largest terrestrial predator, influencing the distribution and abundance of other species, although their precise impact remains debated. Their importance extends to the cultural heritage of Indigenous Australians, featuring prominently in stories and spiritual beliefs. Despite this importance, gaps exist in understanding dingo origins and evolution, including the extent of recent hybridization with domestic dogs and the impact of human activities. Dingoes exhibit variation in appearance across their range, hinting at distinct locally adapted populations. Genetic studies support this, showing southeastern Australian dingoes differ from those elsewhere. Mitochondrial DNA (mtDNA) haplotypes form two clades with geographically exclusive distributions: one in southeastern Australia and K'gari, the other in the rest of the country. Y chromosome and nuclear allele frequencies confirm this division. Some interpret these differences as evidence of at least two independent introductions, though other natural processes may also have contributed. Human activities, such as European colonization and dingo persecution, likely also shaped dingo diversity and distribution. Intriguing similarities exist between dingoes and New Guinea singing dogs, endangered canids with phenotypic resemblance to dingoes. Genetic studies have shown some dingoes have a greater affinity for New Guinea singing dogs, particularly those from southeastern Australia. The timing of gene flow between these canids remains uncertain, hindering understanding of their origins and introduction routes to Australia. The extent of dingo hybridization with domestic dogs is also debated, with implications for conservation. Early studies using microsatellite assays suggested extensive hybridization, but more recent genome-wide SNP analyses suggest hybrid ancestry is rare. All previous studies relied on contemporary dingo populations, lacking a pure baseline for comparison. Paleogenomic data from ancient, pre-Colonial dingoes could resolve these issues. By providing a pre-Colonial snapshot of genetic diversity, these data could establish a minimum age for gene flow between dingoes and New Guinea singing dogs, and the establishment of dingo population structure. This study aims to address these knowledge gaps by sequencing ancient dingo specimens predating European colonization, comparing them to modern dingoes, New Guinea singing dogs, and other canids.

A substantial body of literature exists on dingo origins and evolutionary history. Early studies focused on morphological comparisons, highlighting variations across the dingo's range (Corbett, 1985). These morphological differences were later corroborated by genetic studies, initially using mitochondrial DNA (mtDNA) analyses (Savolainen et al., 2004; Oskarsson et al., 2011), which revealed deep splits between mtDNA haplotypes and suggested multiple introductions to Australia. Subsequent studies incorporated nuclear DNA data, further refining our understanding of dingo population structure and the existence of distinct lineages (Cairns and Wilton, 2016; Cairns et al., 2017; Cairns et al., 2018). The relationship between dingoes and New Guinea singing dogs has also been a topic of ongoing investigation. Initial studies based on mtDNA highlighted genetic similarities, particularly between southeastern Australian dingoes and New Guinea singing dogs (Cairns and Wilton, 2016; Surbakti et al., 2020). The question of dingo-domestic dog hybridization has been particularly contentious. Earlier studies using microsatellite markers suggested widespread hybridization (Stephens et al., 2015), while more recent studies employing genome-wide SNP data challenged this conclusion, indicating a lower prevalence of hybridization (Cairns et al., 2023). The availability of ancient DNA has significantly advanced the field, allowing for investigations into pre-Colonial dingo populations. Studies utilizing ancient canine DNA have provided crucial insights into the broader history of dog domestication (Bergström et al., 2020; Bergström et al., 2022; Ni Leathlobhair et al., 2018). This study builds upon these existing studies by utilizing ancient dingo genomes to directly address the outstanding questions surrounding dingo origins, population structure, and the extent of hybridization.

This study analyzed genomic data from 42 ancient dingo specimens, collected from various sites across Australia, including coastal New South Wales, the Nullarbor Plain, and Western Australia. The majority of specimens predate European colonization, offering a unique insight into pre-Colonial genetic diversity. DNA was extracted and subjected to various processing steps, including library preparation and whole genome enrichment (for some specimens). Whole genome hybridization enrichment was conducted using two different bait sets: one using modern domestic dog DNA and another based on modern K'gari dingo genomic DNA. The processed libraries were sequenced using Illumina platforms (HiSeq 4000, HiSeq X, NextSeq 500), generating both mitochondrial and nuclear genomic data. Ancient DNA authenticity was verified through DNA damage profiling and contamination assessment. Quality control measures were implemented throughout the data processing pipeline. A total of 16 ancient dingo specimens yielded sufficient mitochondrial DNA data, while nine had sufficient nuclear genomic data for further analysis. This ancient genomic data was compared to data from three modern K'gari dingoes (considered to have minimal modern dog ancestry) and previously published data from other modern dingoes from across Australia, New Guinea singing dogs, and other domestic dogs worldwide. Data analysis included mitochondrial DNA haplotype network construction using PopART and BEAST for time-scaled phylogeny reconstruction. Principal Component Analysis (PCA) was performed using Eigensoft SmartPCA to visualize relationships between ancient and modern canids. qpWave analyses and hierarchical clustering investigated ancient dingo sample relationships. f-statistics (f3, f4) were employed to test for allele sharing patterns, admixture events, and assess the proportion of dingo ancestry in modern dingoes. qpGraph was used to model relationships between ancient dingoes, New Guinea singing dogs, and other canids, and DATES assessed the timing of admixture pulses. Microsatellite assay data from previous studies was used to compare ancestry estimates from microsatellite and genome-wide data.

The study's key findings revolve around the ancient history of dingo populations and their relationships to other canids. Mitochondrial DNA (mtDNA) analysis revealed two distinct clusters of dingo haplotypes, mirroring the previously identified northwest and southeast lineages. Ancient dingo mtDNA haplotypes were found to be nested within these modern lineages, demonstrating continuity in maternal ancestry. The estimated time to most recent common ancestor (TMRCA) for these lineages ranged from approximately 3,000 to 8,000 years before present (BP), providing a timeframe for dingo introduction and early diversification. Nuclear genome analysis largely confirmed these mtDNA findings, showing strong affinities between ancient and modern dingoes within the respective lineages. Interestingly, ancient dingoes from coastal New South Wales displayed a significantly greater affinity for New Guinea singing dogs compared to ancient dingoes from the Nullarbor. This asymmetry in allele sharing suggests a closer relationship between ancient eastern Australian dingoes and New Guinea singing dogs. This signal is not solely attributable to ascertainment bias from whole-genome enrichment, as evidenced by the stronger affinity between ancient NSW dingoes and New Guinea singing dogs compared to their affinity with K'gari dingoes. Demographic modeling, utilizing qpGraph, indicated that models incorporating admixture between ancient coastal NSW dingoes and New Guinea singing dogs were superior. The estimated timing of this admixture event was around 2,456 ± 171 BP, consistent with the estimated TMRCA for relevant mtDNA haplotypes. Importantly, the analysis of modern dingoes previously identified as pure using microsatellite assays revealed minimal to no post-Colonial domestic dog ancestry in most individuals. Exceptions, such as a sample from the Gibson Desert and an alpine dingo, highlight the potential for underpowered detection of hybridization in some assays. This underscores the importance of utilizing ancient genomes as a baseline for accurate assessment of dingo ancestry.

The findings provide substantial new insights into dingo evolutionary history. The ancient genomic data confirmed the previously observed population structure in modern dingoes, indicating its deep temporal roots. The excess allele sharing between ancient coastal NSW dingoes and New Guinea singing dogs suggests several possible scenarios: gene flow between ancestral populations outside of Australia, movement of New Guinea singing dogs to eastern Australia and interbreeding with local dingo populations, or the introduction of a single ancestral population to Australia that subsequently differentiated and shared alleles with New Guinea singing dogs. These scenarios aren't mutually exclusive, and further research is needed to fully resolve the complex relationships between these populations. The estimated timing of admixture around 2,456 BP suggests human involvement in facilitating gene flow between Australia and New Guinea, possibly through limited cultural interaction or trade. The consistency in dingo ancestry across most modern dingoes previously designated as 'pure' indicates limited post-colonial hybridization, supporting recent genome-wide findings. The exceptions highlight limitations in earlier microsatellite-based purity assessments. These results underscore the importance of integrating ancient DNA data into future dingo conservation and management strategies.

This study provides compelling evidence for the deep temporal roots of dingo population structure and demonstrates the value of ancient DNA in clarifying previously contentious issues surrounding dingo origins and hybridization. The study confirms the strong affinities between ancient and modern dingoes and reveals a closer relationship between ancient east coast dingoes and New Guinea singing dogs. Further research incorporating ancient genomes from Island Southeast Asia and New Guinea will be crucial in refining our understanding of dingo origins and their relationships with other canid populations. Future efforts should focus on acquiring additional ancient genomic data from relevant geographic regions to further refine the proposed models and resolve the ambiguities surrounding the direction and extent of gene flow between dingoes and New Guinea singing dogs.

While this study significantly advances our understanding of dingoes, certain limitations exist. The limited number of ancient dingo genomes with high-quality data restricts the power of some analyses. Furthermore, biases related to sample preservation and DNA extraction techniques could affect the results. The study primarily focuses on mitochondrial and nuclear genomic data, and integration of other data types (e.g., Y chromosome data) could further enhance the analyses. The interpretation of admixture events relies on modeling approaches, which have inherent uncertainties, highlighting the need for continued investigation and refinement of these models.