Abstract
In many endangered species with increased homozygosity, population recovery is limited by expressed deleterious mutations. In contrast, the critically endangered red-headed wood pigeon, endemic to an isolated oceanic archipelago of ~100 km2, experienced a dramatic population increase three years after the removal of introduced predators, despite being nearly extinct in the 2000s. To understand the reason for this recovery, we sequenced and analysed its genome. We find that over 80% of the genome is homozygous due to centuries of inbreeding, yet the frequency of highly deleterious mutations is lower than in widespread conspecifics. Linkage disequilibrium-based analysis suggests that the effective population size declined from several thousand to fewer than a hundred in the early 20th century due to human impacts. Gradual inbreeding in a historically small population before human impact may have purged part of the mutation load, facilitating the recovery. Our study highlights the role of historical demography in the persistence of bottlenecked populations, with implications for biological conservation.
Similar content being viewed by others
Introduction
Habitat destruction, degradation and overexploitation reduce population sizes and increase the extinction risk of species. Even after such external threats are removed through conservation efforts, internal factors associated with small population size often limit the recovery of endangered populations (e.g., the Florida panther1; the grey wolf2). Some species may become extinct within a short time despite intensive conservation measures, as seen in cases such as the Ogasawara holly blue3. One of the major internal threats is the expression of deleterious mutations4. Recessive or partially recessive deleterious mutations are thought to exist in the genomes mainly in a heterozygous state. When population size decreases and homozygosity increases, these mutations can become homozygous and expressed, leading to reduced fitness5. The total amount of both expressed and masked deleterious mutations in a population is referred to as the genetic load, which is defined as a function of the deleteriousness of alleles (i.e., selection coefficients), their frequencies in the population, and the number of loci carrying such alleles6,7. Genetic load can reduce fitness across current and future generations, thereby hindering population recovery and sometimes accelerating population decline (i.e., the extinction vortex8). However, some endangered species seem to have high recovery potential. For example, the northern elephant seal9, Island fox10, Seychelles paradise flycatcher11, and red-headed wood pigeon12 were once driven to the brink of extinction, but their wild populations increased shortly after the removal of external threats, even without intensive human intervention. These cases raise the question of whether such species carried a relatively low genetic load. Understanding the accumulation and impact of genetic load could help explain why endangered species have different recovery potentials.
Genetic load in a population is shaped by historical changes in effective population size (i.e., population dynamics). In historically small populations, recessive deleterious mutations are more likely to become homozygous and expressed, and thereby more likely to be removed by purifying selection. These mutations are also more likely to be lost or fixed due to strong genetic drift. Over generations, these processes can act together to shape the overall genetic load. In particular, in historically small populations, deleterious mutations are expected to have been effectively purged by selection and drift, resulting in a lower genetic load compared to larger populations7,13,14. Consequently, even if the population size is further reduced by anthropogenic impacts, the resulting inbreeding may be less likely to cause a fitness decline15. If the population remains very small (Ne < ~25) after a bottleneck, the stochastic effects of genetic drift are predicted to facilitate the fixation of deleterious mutations5,7, eventually leading to a mutation meltdown16. However, empirical studies on the relationship between historical population dynamics and genetic load (or the extent of purging) are limited, particularly in successfully recovering endangered species10,11,17,18.
One of the critically endangered species that is steadily recovering through conservation efforts is the red-headed wood pigeon (Columba janthina nitens) (Fig. 1). This bird is endemic to the Ogasawara Islands, an oceanic archipelago about 1000 km from the Japanese main island, Honshu. The total area of the islands is 106.1 km2, of which 69.82 km2 is forested. The Ogasawara Islands were listed as a UNESCO World Natural Heritage site in 2011 for being an outstanding example of the active and ongoing evolutionary processes in oceanic island ecosystems, characterised by the exceptionally high rates of endemism in land snails ( > 90%) and vascular plants (37%). The red-headed wood pigeon favours mature forests and is genetically and ecologically differentiated from the Japanese wood pigeon (C. j. janthina), a subspecies widely distributed on islands near the main Japanese islands and the Korean peninsula19. After the red-headed wood pigeon diverged from the Japanese wood pigeon 0.67 million years ago, the population size has remained orders of magnitude smaller than that of the Japanese wood pigeon20 (Fig. 1). The Ogasawara Islands have been inhabited by humans since 183021. Due to human impact, including forest destruction and predation by introduced feral cats, the red-headed wood pigeon population declined to fewer than 80 individuals in 2008 and was close to extinction22. Intensive trapping of feral cats began in 2010 on Chichijima Island in the Ogasawara Islands. As a result, 131 feral cats were trapped between 2010 and 2013, and the cat population was reduced to less than 20 individuals. Over the same period, the total number of pigeons observed increased from 111 to 966 adults and from 9 to 189 juveniles12. The ex-situ conservation population was established in 2001 with three founders at the Ueno Zoological Gardens, Tokyo. Initially, breeding was repeated between the founders and their offspring, resulting in increased inbreeding in captivity. Since 2012, thirteen wild individuals have been introduced into the captive population. By 2022, the fifth generation had been born, and a total of 198 individuals had been kept in captivity.
We hypothesise that long-term isolation on small oceanic islands has facilitated the purging of genetic load in the red-headed wood pigeon. To gain insight into the reasons for this species’ recovery, we analyse genomic, pedigree, and fitness data and address the following three objectives. (1) To characterise the level of inbreeding and genetic load and to assess the impact of historical population dynamics on genetic load in the red-headed wood pigeon, we compare the extent of runs of homozygosity (ROH) and nonsense single nucleotide polymorphisms (SNPs) in its genome with those in the captive population and in the widely distributed Japanese wood pigeons. (2) To quantify the genetic load in the red-headed wood pigeon, we examine inbreeding depression for survival in the captive population. (3) To characterise the timing and magnitude of population bottlenecks in the red-headed wood pigeon, we estimate changes in effective population size over the past centuries.
Results
Resequencing and genotyping
We sequenced 25 genomes, including 8 wild red-headed wood pigeons, 8 captive red-headed wood pigeons and 9 wild Japanese wood pigeons. After adapter trimming, an average of 97.8% of reads (standard deviation [SD] = 1.3) was mapped to the rock pigeon (Columba livia) assembly23, and an average of 97.5% (SD = 0.5) of the assembly was covered by at least five reads. The mean coverage per sample was 41.0× (SD = 12.2), and a total of 7,033,246 autosomal SNPs were detected across the samples. Details of the samples and data quality are shown in Supplementary Data 1.
Inbreeding coefficient estimated from ROH
Individual inbreeding coefficient (FROH) was estimated as the proportion of the autosomes covered by runs of homozygosity (ROH) longer than 0.1 Mb. ROH refers to genomic regions with almost no heterozygous sites over a given length, formed by the autozygosity of copies of an ancestral chromosome24. Ancestral chromosome segments become fragmented by recombination, and the length of an ROH formed by two copies that have each experienced m recombination events is expected to be 100/2 m cM25,26. Assuming a genome-wide average recombination rate of 2 cM/Mb and a generation time of 3.58 years in the red-headed wood pigeon27, a 0.1 Mb ROH is expected to derive from roughly 895 years ago. The abundance of ROH in this length category therefore suggests a small effective population size during that time period26.
Overall, FROH was significantly higher in red-headed wood pigeons than in Japanese wood pigeons, and longer ROH were more abundant in the captive than in the wild red-headed wood pigeons (Fig. 2a, b). The mean FROH was 0.02 (range: 0.01–0.03) in Japanese wood pigeons, 0.84 (range: 0.83–0.86) in the wild red-headed wood pigeons, and 0.88 (range: 0.87–0.92) in the captive individuals (Fig. 2b). Long ROH larger than 1 Mb (roughly originating 90 years ago) were found in all the wild and captive red-headed wood pigeons, but were absent in one Japanese wood pigeon (Fig. 2b). The long ROH comprised an average of 20.2% (range: 14.3–26.9%) of the genome in the wild red-headed wood pigeons and 40.4% (range: 32.9–58.6%) in the captive individuals. Very long ROH larger than 10 Mb (roughly originating 895 years ago) were detected in all the captive red-headed wood pigeons, whereas it was found in only three of eight wild red-headed wood pigeons and was not observed in any Japanese wood pigeons (Fig. 2b). The very long ROH comprised an average of 2.03% (range: 1–8%) of the genome in the wild red-headed wood pigeons and 11.2% (range: 1–30%) in the captive individuals.
a Distribution of runs of homozygosity (ROH) longer than 1 Mb and nonsense single nucleotide polymorphisms (SNPs) on the genome. Each line represents an individual (e.g., NA030). For clarity, only the 10 largest autosomes are shown here; the analysis covered all 38 autosomes. b Individual inbreeding coefficients (FROH), calculated as the fraction of the genome in ROH regions longer than 0.1, 1, or 10 Mb. c Allele frequency of each nonsense SNP. Grey lines connect six SNPs found in all three populations. d Proportion of nonsense effects relative to silent effects of derived alleles for each sample.
Deleterious mutations
The abundance of nonsense SNPs in the genome was used as a proxy for the relative genetic load among populations. To identify nonsense mutations, we used a predictive approach based on gene annotation data. We predicted the effect of derived alleles on protein coding using SnpEff version 5.128, assuming that the reference allele is ancestral and the alternative is derived. For the red-headed wood pigeon, 13 nonsense SNPs were detected in both the wild and captive individuals, with 8 of these being nearly fixed in each population (allele frequency >0.9) (Fig. 2a, c; Supplementary Data 2). In contrast, 96 nonsense SNPs were found in the Japanese wood pigeons. One of them was nearly fixed, but this SNP was also fixed in the red-headed wood pigeons. Of the six nonsense SNPs found in all three populations, four were at low frequencies in the Japanese wood pigeons but fixed in the red-headed wood pigeons (Fig. 2c). The ratio of nonsense to silent effects was significantly lower in the wild and the captive red-headed wood pigeons than in the Japanese wood pigeons (ANOVA: F2,22 = 10.43, p = 0.00065; Tukey’HSD: wild red-headed wood pigeons vs. Japanese wood pigeons, p = 0.0000044; captive red-headed wood pigeons vs. Japanese wood pigeons, p = 0.0000011; Fig. 2d). No significant difference was found between the wild and the captive red-headed wood pigeons (p = 0.82).
Inbreeding depression
To quantify the genetic load in the red-headed wood pigeon, we examined inbreeding depression for survival in 119 captive-born individuals with known inbreeding coefficients and longevity (in days). A linear regression analysis showed no significant effect of the inbreeding coefficient on longevity (F1,117 = 2.73, p = 0.101; Fig. 3). While not statistically significant, the positive slope of the regression line (β = 3.75; 95% confidence interval [CI]: −0.70 to 8.19) implies that longevity increased with the level of inbreeding.
The solid line represents the fitted linear regression model (F1,117 = 2.73, p = 0.101). The estimated slope was β = 3.75 (95% CI: −0.70 to 8.19).
Recent population size
We estimated recent changes in the effective population size (Ne) of the red-headed wood pigeon based on linkage disequilibrium data on macrochromosomes using GONE29. Since the recombination rate in this species is unknown, we repeated the inference assuming recombination rates of either 1 or 2 cM/Mb, which span the typical range of average recombination rates for macrochromosomes in other bird species (Supplementary Table 1). The inferred temporal trends in Ne were similar between the two assumptions, but the absolute estimates varied depending on the assumptions (Fig. 4). Assuming a recombination rate of 1 cM/Mb, Ne was estimated to have declined sharply from a mean of 8204 (95% CI: 6523–10,318) in 1892 to a minimum of 88 (95% CI: 84–93) in 1957, followed by an increase to 498 (95% CI: 448–554) in 2007. In contrast, under the assumption of 2 cM/Mb, Ne was estimated to have declined from 1382 (95% CI: 1172–1629) in 1957 to a minimum of 56 (95% CI: 51–63) in 1985, and then increased to 117 (95% CI: 107–128) in 2007.
Ne was estimated from linkage disequilibrium on macrochromosomes of 8 individuals sampled between 2001 and 2021. The orange ribbon represents the 95% confidence interval based on 100 bootstrap replicates, and the orange line shows the geometric mean. Estimates were repeated assuming average recombination rates of 1 and 2 cM/Mb on macrochromosomes. The black line represents changes in the human population of the Ogasawara Islands. The generation time of 3.58 years was used to convert generations into years. Note that estimates during the sampling period can be less accurate29.
Discussion
The red-headed wood pigeon (Columba janthina nitens) is an endemic species of the Ogasawara Islands, an oceanic archipelago inhabited by humans since 183021. By 2008, its population had declined to fewer than 80 individuals due to forest destruction and predation by introduced feral cats22. Three years after the removal of these predators, the population increased markedly12. To understand the reason for this recovery, we sequenced and analysed its genome to infer past demographic trajectories and estimate the population’s genetic load. Our analysis revealed the recent population bottleneck that occurred between the 1900s and 1950s, a period of elevated human population on the islands (Fig. 4). Despite over 80% of the genome being homozygous due to inbreeding, the species carried fewer deleterious mutations than its widespread conspecific, the Japanese wood pigeon (C. j. janthina) (Fig. 2). Here, we discuss how the pigeon’s past demographic history shaped its genetic load and contributed to the population recovery, and we consider the implications for its conservation.
The GONE analysis estimated that the recent population bottleneck began between the 1900s and 1950s, a period of elevated human population in the Ogasawara Islands (Fig. 4). In the early 20th century, the human population growth led to widespread deforestation for agriculture30. Notably, in 1945, the human population reached its peak due to World War II, with around 43,700 soldiers in the islands31, suggesting intensive deforestation and overexploitation at that time30. The GONE result thus indicates that human activities likely triggered the recent bottleneck in the red-headed wood pigeon. The estimated Ne before the recent bottleneck (1382–8204 individuals; Fig. 4) is similar to the size inferred ~10,000 years ago by the PSMC method in a previous study20 (~8000 individuals; Fig. 1), suggesting that the red-headed wood pigeon maintained an Ne of several thousand on these oceanic islands until the recent bottleneck. Although the GONE analysis inferred a post-bottleneck increase in Ne, it remains unclear whether this trend reflects an actual increase in census size. This is because recovery in census size has only been observed since the 2010s12, and GONE estimates can be less accurate during the sampling period29.
Some wild red-headed wood pigeons had very long ROH > 10 Mb, as did the highly inbred captive individuals (Fig. 2b). This suggests that wild red-headed wood pigeons experienced recent inbreeding between closely related individuals who shared a common ancestor a few generations ago. In wild red-headed wood pigeons, over 80% of the genome was in ROH > 0.1 Mb (Fig. 2b). Given that a 0.1 Mb ROH is expected to have originated from a common ancestor roughly 895 years ago, this high proportion of short ROH indicates that the red-headed wood pigeon population was already small before human arrival in 1830 (190 years ago). The fraction of the genome in ROH > 0.1 Mb is higher in red-headed wood pigeons than in other critically endangered birds (kākāpō32, Chatham Island black robin33), although the ROH detection method varies among studies. Thus, the ROH data support increased inbreeding in wild red-headed wood pigeons. The significantly higher abundance of very long ROH > 10 Mb in captive red-headed wood pigeons (Fig. 2b) supports fast inbreeding between closely related birds in captivity. The near absence of ROH longer than 0.1 Mb in the Japanese wood pigeons (Fig. 2b) supports our assumption that their population size has been relatively large.
The red-headed wood pigeon had a lower ratio of nonsense to silent mutations (Fig. 2c) compared to the Japanese wood pigeon, which has historically maintained a larger population size20 (Fig. 1). This suggests that deleterious mutations were more effectively purged in the red-headed wood pigeon. Purging of deleterious mutations is generally more effective under slow inbreeding than under fast inbreeding34. Therefore, we propose that this effective purging was facilitated during the long-term persistence at the relatively small population size prior to the recent bottleneck. Gradual inbreeding over a long period on the small oceanic islands may have reduced the genetic load in the red-headed wood pigeon compared to that of the Japanese wood pigeon. We also found no significant difference in the ratio of nonsense to synonymous mutations between the wild and captive populations of the red-headed wood pigeons (Fig. 2d). This result is consistent with the previous hypothesis that purging is less effective under rapid inbreeding, such as that which may occur in captivity35. In addition to purging, many deleterious mutations were nearly fixed in the red-headed wood pigeon due to its historically small population size (Fig. 2c). While these fixed mutations no longer contribute to inbreeding depression, they could still reduce the overall fitness of the red-headed wood pigeon population. Nevertheless, this potential reduction in fitness does not appear to have critically affected the species’ viability, as evidenced by the recent recovery in the census size.
The positive slope of the regression line in Fig. 3 suggests that longevity tended to increase (though non-significantly) with the level of inbreeding. Explaining this pattern in terms of genetic load is difficult, as it would imply the birds carried a negative number of lethal equivalents—a biologically implausible scenario. Rather, we propose that this trend might be due to advancements in captive breeding practices over time. While inbreeding has increased across generations in captivity, hand-rearing techniques, for instance, have also improved. This temporal coincidence could have led to the spurious positive correlation between inbreeding and longevity. It should also be noted that the genetic load could be underestimated due to the following three potential limitations in the estimation: (1) Inbreeding depression for reproduction is not evaluated in this study, owing to potential bias in fertility data caused by captive management. (2) Hatching success, which is known to be affected by inbreeding in other birds36,37, was not assessed due to the lack of data. (3) Inbreeding depression may be less pronounced in captivity, where the environmental conditions are generally milder than in the wild38.
Purging of deleterious mutations in endangered species recovering from near extinction has also been reported for the Island fox10, Seychelles paradise flycatcher11, northern elephant seal39, and Iberian lynx18. These species, like the red-headed wood pigeon, have experienced long-term isolation with relatively small population sizes. A contrasting case is the pink pigeon, an endangered species endemic to the oceanic island of Mauritius, which had a relatively large ancestral population size (Ne = ~16,000). Its wild population increased from a bottleneck through intensive management, including the release of captive-born birds, but its population growth has since stalled40. The pink pigeon carries a substantial genetic load of 15.13 LEs for longevity and therefore faces a high extinction risk40,41. These contrasting cases suggest that historical population dynamics prior to human impact can influence the recovery potential of endangered populations. For the red-headed wood pigeon, the relatively short duration of the bottleneck, particularly the period with Ne < ~100, may also have contributed to its recovery by limiting the accumulation of deleterious mutations due to genetic drift.
Our findings have several implications for the conservation of the red-headed wood pigeon. The purging of deleterious mutations (Fig. 2c, d) and the lack of a significant relationship between inbreeding and longevity (Fig. 3) suggest that inbreeding has a limited impact on fitness and that this species has a relatively high recovery potential from a bottleneck. Still, this does not necessarily ensure its long-term viability. The low genetic diversity may reduce adaptability to environmental changes, thereby compromising long-term persistence42, as shown for the Seychelles paradise flycatcher11. Moreover, if the population has not yet reached genetic equilibrium after the recent bottleneck, further loss of genetic diversity and fixation of deleterious mutations may still occur43,44, as seen in the case of the pink pigeon40,41. If the population does not recover sufficiently in the future, genetic drift may further erode diversity and promote the fixation of deleterious alleles5,7,41. Mutation accumulation may also be facilitated by relaxed selection under intensive conservation management45. Therefore, increasing the size of the captive population, alongside continued genetic monitoring and management, will be essential for the long-term persistence of the red-headed wood pigeon. As illustrated by the pink pigeon, incorporating genetic load data into ex situ conservation strategies can help reduce genetic load while maintaining genetic diversity46.
The Ogasawara Islands host hundreds of endemic plant species, while there are only a few seed-dispersing bird species, such as pigeons. The endemic Bonin wood pigeon (Columba versicolor) went extinct after 1889 due to human impacts. The red-headed wood pigeon is today one of the few large seed dispersers remaining on the Ogasawara Islands. Therefore, its persistence will be crucial for sustaining the distinctive ecosystem with such a high diversity of endemic plants. The present study illustrates the purging of highly deleterious mutations in inbred island endemics recovering from near extinction. Our findings provide genomic evidence for the hypothesis that species with long-term, small population sizes can become tolerant to inbreeding, thus deepening our understanding of the factors contributing to the persistence of small populations. We also present individual-level information on the genome-wide distribution of eroded regions, including ROH and deleterious mutations, in the critically endangered species (Fig. 2a). This will assist in the selection of captive population founders, breeding pairs in captivity, and individuals for genetic rescue, thereby facilitating the future advancement of more informed and effective conservation strategies.
Methods
Genomic sample collection
We collected genomic samples from three populations (wild and captive populations of the red-headed wood pigeon and the wild population of the Japanese wood pigeon) to compare their genomes. We collected nine wild red-headed wood pigeons rescued between 2001 and 2021 on Chichijima Island. We collected nine captive red-headed wood pigeons that were born at Ueno Zoological Gardens and Tama Zoological Park between 2008 and 2018 and had especially high inbreeding coefficients due to close inbreeding in captivity (pedigree-based inbreeding coefficients = 0.25–0.36; Supplementary Fig. 1). We also obtained eleven wild Japanese wood pigeons collected between 1995 and 2017 and archived at the Yamashina Institute for Ornithology, Japan. Sample details are shown in Supplementary Data 1. This study did not involve any animal experimentation, anaesthesia, or euthanasia. All samples were obtained either non-invasively (e.g., naturally shed feathers or tissues) or from previously extracted DNA stored in museum collections. Therefore, no institutional ethical approval (e.g., IACUC) was required.
DNA extraction, library preparation, and sequencing
DNA was extracted from feathers, liver, or blood of each individual using QIAGEN DNeasy Blood and Tissue Kit. Genomic libraries were prepared with Illumina DNA Prep Kit following the manufacturer’s protocol, and sequenced with 150 bp paired-end reads on one lane of the BGI DNBSEQ-T7.
Sequence read processing and mapping
Adapter sequences in raw reads were removed with Trimmomatic version 0.3947. As a reference sequence, we used an assembly of the rock pigeon (Columba livia)23. Two chromosomes in the assembly (Genbank: CM053661.1, CM053672.1) were excluded from downstream analyses because they showed high similarity to the Z chromosome (LR594555.2) of turtle dove (Streptopelia turtur) in our genome comparison using D-GENIES48. Repeat sequences in the assembly were masked using RepeatModeler version 2.0.449 and RepeatMasker version 4.1.550. Processed reads were mapped to the reference using BWA-MEM2 version 2.2.151, and duplicate reads were marked using samtools version 1.1752. Two samples (NA031, NA032) were excluded because less than 50% of the reads were successfully mapped. Details of the sequencing data quality are provided in Supplementary Data 1.
Variant calling and filtering
SNPs were detected for each sample using bcftools version 1.10.253. To obtain high-quality data, we selected genotype data meeting the following criteria: (1) depth of at least 5, greater than one-third of the mean depth, and less than twice the mean depth; (2) allelic balance greater than 0.2 and less than 0.8 at heterozygous sites; (3) genotype quality higher than 20; and (4) more than 10 bp away from indels. The per-sample variant data (VCF files) were merged, and the proportion of missing data was calculated for each sample. We excluded two samples (NA004, NA020) with significantly high missing data (Supplementary Fig. 2) and retained 25 samples. We selected SNP sites meeting the following criteria: (1) sites with no missing data; (2) single alternative allele (i.e., biallelic sites); and (3) heterozygosity less than 0.9. Finally, we obtained 7,033,246 SNP loci.
Inbreeding coefficient estimated from ROH
Individual inbreeding coefficients were estimated from runs of homozygosity (ROH). ROH are genomic regions with almost no heterozygous sites over a given length and are formed by the autozygosity of copies of an ancestral chromosome24. The inbreeding coefficient (FROH) was calculated as FROH = total length of ROH (S) / total length of genome (L)54. ROH were detected using bcftools roh55. The total length of ROH (S) was calculated by varying the different minimum ROH length to 0.1, 1 and 10 Mb.
ROH cannot be detected in large gaps of an assembly or in genomic regions with low marker density. In such cases, it is not appropriate to use the total length of the autosomal genome as the total length of the genome (L). To account for these regions, we simulated an individual with a completely homozygous genotype as described in Meyermans et al.56. The total length of the ROH detected in the homozygous individual was considered the maximum detectable ROH and used as the total length of the genome (L). L was 1,002,121,181 bp, which means that we could detect ROH in almost all (99.9%) regions of the autosomal genome.
Deleterious mutations
To identify deleterious mutations, we employed a predictive approach based on gene annotation data to ascertain the effects of mutations on protein coding. Since the reference assembly lacked gene annotation, the gene annotation of a scaffold-level assembly of the rock pigeon (GenBank: GCA_000337935.2) was mapped to the reference using Liftoff version 1.6.357. A total of 14,374 out of 15,392 genes (93.4%) were successfully mapped. Assuming that the reference allele is ancestral and the alternative allele is derived, we predicted the effect of the derived allele on protein coding using SnpEff version 5.128. To ensure the accuracy of the predictions, we excluded those with a ‘warning’ message, which indicates a possible error in the reference genome. The relative number of nonsense to silent mutations was calculated for each sample, and differences among populations were tested using ANOVA followed by Tukey’s HSD.
Inbreeding depression
We quantified the genetic load in the red-headed wood pigeon by examining inbreeding depression for survival in the captive population. The level of inbreeding depression is defined as −B, where B is the slope of a linear regression of the natural logarithm of longevity on the inbreeding coefficient4,6,7. The genetic load in diploid organisms is then defined as −2B, expressed in units of lethal equivalents. Of the 197 birds listed in the studbook with known parentage, we analysed 119 captive-born individuals that died between July 2003 and January 2022. The inbreeding coefficient was calculated from the pedigree using the R package pedigreemm 0.3.458. Longevity was calculated as the number of days from hatching to death. We performed a generalized linear model (GLM) analysis using the inbreeding coefficient as the explanatory variable and the natural logarithm of longevity plus one, i.e. log(longevity + 1), as the response variable. Inbreeding depression for reproduction was not evaluated, owing to potential bias in fertility data caused by captive management.
Recent population size
We estimated recent changes in the effective population size (Ne) of the red-headed wood pigeons based on linkage disequilibrium between SNPs using GONE29. Since microchromosomes may not retain signals of past demographic events due to their particularly high recombination rate, we restricted our analysis to macrochromosomes (the ten autosomes larger than 40 Mb; Supplementary Fig. 3). We detected biallelic SNPs from the eight samples of wild red-headed wood pigeons and retained only sites with no missing data. The hc parameter was set to 0.05 according to the developer’s recommendation. A hundred bootstrap replicates were performed by randomly sampling up to 30,000 loci from each chromosome. To convert generations to years, we used a generation time of 3.58 years for this species22. Because the recombination rate of this species is unknown, we referred to estimates from other bird species (Supplementary Table 1). Given that the average recombination rate on macrochromosomes typically range from 1 to 2 cM/Mb, we repeated the inference assuming recombination rates of either 1 or 2 cM/Mb. To assess the robustness of the inference given the small sample size, we repeated the estimation eight times by subsampling seven individuals from the total of eight. The estimates based on seven individuals were consistent with those based on all eight individuals (Supplementary Figs. 4, 5), suggesting that the inference using all eight samples is likely to be reliable. To examine the potential impact of human activity on the pigeon population, we compiled historical human population data of the Ogasawara Islands from the literature31,59,60 (Supplementary Table 2) and compared it with the estimated pigeon population size.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
Whole-genome sequencing data are deposited at DDBJ Sequence Read Archive under the accession numbers DRR599214–DRR599240 (BioProject: PRJDB18884). The analysis scripts and source data for the main figures are available on Dryad repository with the identifier (https://doi.org/10.5061/dryad.j0zpc86q7)61. The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
Roelke, M. E., Martenson, J. S. & O’Brien, S. J. The consequences of demographic reduction and genetic depletion in the endangered Florida panther. Curr. Biol. 3, 340–350 (1993).
Liberg, O. et al. Severe inbreeding depression in a wild wolf Canis lupus population. Biol. Lett. 1, 17–20 (2005).
Nakahama, N. et al. Road to extinction: Archival samples unveiled the process of inbreeding depression during artificial breeding in an almost extinct butterfly species. Biol. Conserv. 296, 110686 (2024).
Frankham, R. Genetics and extinction. Biol. Conserv. 126, 131–140 (2005).
Dussex, N., Morales, H. E., Grossen, C., Dalén, L. & van Oosterhout, C. Purging and accumulation of genetic load in conservation. Trends Ecol. Evol. 38, 961–969 (2023).
Morton, N. E., Crow, J. F. & Muller, H. J. An estimate of the mutational damage in man from data on consanguineous marrages. Proc. Natl Acad. Sci. 42, 855–863 (1956).
Bertorelle, G. et al. Genetic load: genomic estimates and applications in non-model animals. Nat. Rev. Genet. 23, 492–503 (2022).
Gilpin, M. E. & Soulé, M. E. Minimum viable populations: processes of extinction. in Conservation Biology: The Science of Scarcity and Diversity (ed. Soulé, M.E.) 19–34 (Sinauer Associates, Sunderland, MA, 1986).
Abadía-Cardoso, A., Freimer, N. B., Deiner, K. & Garza, J. C. Molecular population genetics of the northern elephant seal Mirounga angustirostris. J. Hered. 108, 618–627 (2017).
Robinson, J. A., Brown, C., Kim, B. Y., Lohmueller, K. E. & Wayne, R. K. Purging of strongly deleterious mutations explains long-term persistence and absence of inbreeding eepression in island foxes. Curr. Biol. 28, 3487–3494 (2018).
Femerling, G. et al. Genetic load and adaptive potential of a recovered avian species that narrowly avoided extinction. Mol. Biol. Evol. 40, msad256 (2023).
Horikoshi, K., Suzuki, H., Sasaki, T. & Kawakami, K. Population increase of the endemic subspecies of the Japanese Wood Pigeon Columba janthina nitens on Chichijima in the Ogasawara Islands, Japan, after feral cat population control. Jpn. J. Ornithol. 69, 3–18 (2020).
Glémin, S. How are deleterious mutations purged? Drift versus nonrandom mating. Evolution 57, 2678–2687 (2003).
Hedrick, P. W. & Garcia-Dorado, A. Understanding inbreeding depression, purging, and genetic rescue. Trends Ecol. Evol. 31, 940–952 (2016).
Kyriazis, C. C., Wayne, R. K. & Lohmueller, K. E. Strongly deleterious mutations are a primary determinant of extinction risk due to inbreeding depression. Evol. Lett. 5, 33–47 (2021).
Lynch, M., Conery, J. & Burger, R. Mutation accumulation and the extinction of small populations. Am. Nat. 146, 489–518 (1995).
Grossen, C., Guillaume, F., Keller, L. F. & Croll, D. Purging of highly deleterious mutations through severe bottlenecks in Alpine ibex. Nat. Commun. 11, 1001 (2020).
Kleinman-Ruiz, D. et al. Purging of deleterious burden in the endangered Iberian lynx. Proc. Natl Acad. Sci. 119, e2110614119 (2022).
Seki et al. Distribution and genetic structure of the Japanese wood pigeon (Columba janthina) endemic to the islands of East Asia. Conserv. Genet. 8, 1109–1121 (2007).
Tsujimoto, D. et al. Has long-distance flight ability been maintained by pigeons in highly insular habitats?. J. Biogeogr. 50, 235–246 (2023).
Hawks, F. L. Narrative of the Expedition of an American Squadron to the China Seas and Japan, Performed in the Years 1852, 1853, and 1854, under the Command of Commodore M.C. Perry, United States Navy, by Order of the Government of the United States. (Beverley Tucker, Senate Printer, Washington, 1856).
Horikoshi, K. et al. Red-Headed Wood Pigeon Population and Habitat Viability Assessment Workshop Final Report. (Conservation Breeding Specialist Group (SSC/IUCN), Apple Valley, MN, 2010).
Wang, Y. et al. Spatio-temporal transcriptome dynamics coordinate rapid transition of core crop functions in ‘lactating’ pigeon. PLOS Genet 19, e1010746 (2023).
Pemberton, T. J. et al. Genomic patterns of homozygosity in worldwide human populations. Am. J. Hum. Genet. 91, 275–292 (2012).
Thompson, E. A. Identity by descent: variation in meiosis, across genomes, and in populations. Genetics 194, 301–326 (2013).
Kardos, M., Qvarnström, A. & Ellegren, H. Inferring individual inbreeding and demographic history from segments of identity by descent in Ficedula flycatcher genome sequences. Genetics 205, 1319–1334 (2017).
Bird, J. P. et al. Generation lengths of the world’s birds and their implications for extinction risk. Conserv. Biol. 34, 1252–1261 (2020).
Cingolani, P. et al. A program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff: SNPs in the genome of Drosophila melanogaster strain w1118; iso-2; iso-3. Fly. (Austin) 6, 80–92 (2012).
Santiago, E. et al. Recent demographic history inferred by high-resolution analysis of linkage disequilibrium. Mol. Biol. Evol. 37, 3642–3653 (2020).
Kawakami, K. The history of anthropogenic disturbance and invasive alien species impact on the indigenous avifauna of the Ogasawara Islands, southern Japan. Jpn. J. Ornithol. 68, 237–262 (2019).
Ogasawara Natural Environmental Study Group. Field guide: Nature in Ogasawara. (Kokinsyoin, Tokyo, 1992).
Dussex, N. et al. Population genomics of the critically endangered kākāpō. Cell Genomics 1, 100002 (2021).
Von Seth, J. et al. Genomic trajectories of a near-extinction event in the Chatham Island black robin. BMC Genomics 23, 747 (2022).
Allendorf, F. W., Funk, W. C., Aitken, S. N., Byrne, M. & Luikart, G. Conservation and the Genomics of Populations. 389–392 (Oxford University Press, Oxford, 2022).
Leberg, P. L. & Firmin, B. D. Role of inbreeding depression and purging in captive breeding and restoration programmes. Mol. Ecol. 17, 334–343 (2008).
Spottiswoode, C. & Møller, A. P. Genetic similarity and hatching success in birds. Proc. R. Soc. Lond. B Biol. Sci. 271, 267–272 (2004).
Heber, S. & Briskie, J. V. Population bottlenecks and increased hatching failure in endangered birds. Conserv. Biol. 24, 1674–1678 (2010).
Jiménez, J. A., Hughes, K. A., Alaks, G., Graham, L. & Lacy, R. C. An experimental study of inbreeding depression in a natural habitat. Science 266, 271–273 (1994).
Hoffman, J. I. et al. Genomic and fitness consequences of a near-extinction event in the northern elephant seal. Nat. Ecol. Evol. 8, 2309–2324 (2024).
Jackson, H. A. et al. Genomic erosion in a demographically recovered bird species during conservation rescue. Conserv. Biol. 36, e13918 (2022).
Pinto, A. V., Hansson, B., Patramanis, I., Morales, H. E. & Van Oosterhout, C. The impact of habitat loss and population fragmentation on genomic erosion. Conserv. Genet. 25, 49–57 (2024).
Willi, Y., Van Buskirk, J. & Hoffmann, A. A. Limits to the adaptive potential of small populations. Annu. Rev. Ecol. Evol. Syst. 37, 433–458 (2006).
Kuussaari, M. et al. Extinction debt: a challenge for biodiversity conservation. Trends Ecol. Evol. 24, 564–571 (2009).
Tilman, D., May, R. M., Lehman, C. L. & Nowak, M. A. Habitat destruction and the extinction debt. Nature 371, 65–66 (1994).
Cavill, E. L. et al. When birds of a feather flock together: Severe genomic erosion and the implications for genetic rescue in an endangered island passerine. Evol. Appl. 17, e13739 (2024).
Speak, S. A. et al. Genomics-informed captive breeding can reduce inbreeding depression and the genetic load in zoo populations. Mol. Ecol. Resour. 24, e13967 (2024).
Bolger, A. M., Lohse, M. & Usadel, B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30, 2114–2120 (2014).
Cabanettes, F. & Klopp, C. D-GENIES: dot plot large genomes in an interactive, efficient and simple way. PeerJ 6, e4958 (2018).
Flynn, J. M. et al. RepeatModeler2 for automated genomic discovery of transposable element families. Proc. Natl Acad. Sci. 117, 9451–9457 (2020).
Smit, A. F. A., Hubley, R. & Green, P. RepeatMasker Open-4.0. (2013).
Vasimuddin, M. D., Misra, S., Li, H. & Aluru, S. Efficient architecture-aware acceleration of BWA-MEM for multicore systems. in IEEE International Parallel and Distributed Processing Symposium 314–324 (IEEE, 2019). https://doi.org/10.1109/IPDPS.2019.00041.
Danecek, P. et al. Twelve years of SAMtools and BCFtools. GigaScience 10, giab008 (2021).
Li, H. A statistical framework for SNP calling, mutation discovery, association mapping and population genetical parameter estimation from sequencing data. Bioinformatics 27, 2987–2993 (2011).
Kardos, M., Luikart, G. & Allendorf, F. W. Measuring individual inbreeding in the age of genomics: marker-based measures are better than pedigrees. Heredity 115, 63–72 (2015).
Narasimhan, V. et al. BCFtools/RoH: a hidden Markov model approach for detecting autozygosity from next-generation sequencing data. Bioinformatics 32, 1749–1751 (2016).
Meyermans, R., Gorssen, W., Buys, N. & Janssens, S. How to study runs of homozygosity using PLINK? A guide for analyzing medium density SNP data in livestock and pet species. BMC Genomics 21, 94 (2020).
Shumate, A. & Salzberg, S. L. Liftoff: accurate mapping of gene annotations. Bioinformatics 37, 1639–1643 (2021).
Vazquez, A. I., Bates, D. M., Rosa, G. J. M., Gianola, D. & Weigel, K. A. Technical note: An R package for fitting generalized linear mixed models in animal breeding. J. Anim. Sci. 88, 497–504 (2010).
Tokyo Prefecture. Overview of the Ogasawara Islands. (Tokyo Prefecture, Tokyo, 1929).
Tokyo Metropolitan Ogasawara Islands Branch Office. Overview of the Administrative Area. (Tokyo Metropolitan Ogasawara Islands Branch Office, Tokyo, 2023).
Tsujimoto, D. et al. Genetic purging in an island-endemic pigeon recovering from the brink of extinction [Data set]. Dryad https://doi.org/10.5061/dryad.j0zpc86q7 (2025).
Acknowledgements
We thank Yamashina Institute for Ornithology for the use of archival tissue samples. We are also grateful to Professor Cock van Oosterhout and two anonymous reviewers for their thoughtful and constructive comments, which greatly improved the clarity and quality of the manuscript. This work was supported by the Environment Research and Technology Development Fund (JPMEERF20224M02) of the Environmental Restoration and Conservation Agency provided by the Ministry of the Environment of Japan.
Author information
Authors and Affiliations
Contributions
D.T. and Y.I conceived and designed the study. Y.I. obtained funding. D.T., M.T., J.I., R.S., K.H. and H.S. acquired the data. D.T. analysed the data. D.T., M.T., J.I., R.S., K.H., H.S. and Y.I interpreted the data. D.T and Y.I. wrote the manuscript with input from all authors.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Communications Biology thanks Cock van Oosterhout, Oliver Ryder and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary Handling Editors: Michele Repetto and Aylin Bircan. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Tsujimoto, D., Takayanagi, M., Ishii, J. et al. Genetic purging in an island-endemic pigeon recovering from the brink of extinction. Commun Biol 8, 1051 (2025). https://doi.org/10.1038/s42003-025-08476-z
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s42003-025-08476-z
