Genomic consequences of isolation and inbreeding in an island dingo population

Abstract

Dingoes come from an ancient canid lineage that originated in East Asia around 8000-11,000 years BP. As Australia's largest terrestrial predator, dingoes play an important ecological role. A small, protected population exists on a world heritage listed offshore island, K'gari (formerly Fraser Island). Concern regarding the persistence of dingoes on K'gari has risen due to their low genetic diversity and elevated inbreeding levels. However, whole-genome sequencing data is lacking from this population. Here, we include five new whole-genome sequences of K'gari dingoes. We analyze a total of 18 whole genome sequences of dingoes sampled from mainland Australia and K'gari to assess the genomic consequences of their demographic histories. Long (>1 Mb) runs of homozygosity (ROH) - indicators of inbreeding - are elevated in all sampled dingoes. However, K'gari dingoes showed significantly higher levels of very long ROH (>5 Mb), providing genomic evidence for small population size, isolation, inbreeding, and a strong founder effect. Our results suggest that, despite current levels of inbreeding, the K'gari population is purging strongly deleterious mutations, which, in the absence of further reductions in population size, may facilitate the persistence of small populations despite low genetic diversity and isolation. However, there may be little to no purging of mildly deleterious alleles, which may have important long-term consequences, and should be considered by conservation and management programs.

Fig. 1.

Reduced dimensional representation of sample clustering based on multidimensional scaling of the allele sharing dissimilarity matrix. First two major axes of variation plotted for (A) all 41 canids included in this study and (B) 18 dingo individuals.

Fig. 2.

Genetic diversity estimates within and between canid populations. Error bars representing 95% confidence intervals for each line are too narrow to be visible. (A) Mean number of alleles per locus (allelic richness). (B) Mean number of private alleles per locus (private allelic richness). (C) Mean number of shared alleles per locus between pairs of populations.

Fig. 3.

ROH coverage across individual genomes. The percentage of the genome covered by class A ROH (0.5–1 Mbp), class B ROH (1–3 Mbp), class C ROH (3–5 Mbp) and class D ROH (>5 Mbp) is shown on the Y-axis for each of the 41 canids included in this study.

Fig. 4.

Total number of predicted damaging genotypes in each canid population. (A) Alternate-allele deleterious homozygotes. (B) Alternate-allele deleterious heterozygotes. (C) Alternate-allele LoF homozygotes. (D) Alternate-allele LoF heterozygotes. P-value significance by Wilcoxon test with K’gari dingoes as reference population. Significance levels of p-values: (*) p-value < 0.05, (**) p-value < 0.005, (ns) p-value > 0.05.

Fig. 5.

Kinship-weighted Rxy estimates for K’gari and mainland dingoes. Estimates represent the relative excess number of derived alleles for each category in K’gari dingoes (population X) relative to mainland dingoes (population Y). Error bars represent the 95% confidence interval from jackknife resampling.