Genomic trajectories of a near-extinction event in the Chatham Island black robin

Abstract

Background: Understanding the micro--evolutionary response of populations to demographic declines is a major goal in evolutionary and conservation biology. In small populations, genetic drift can lead to an accumulation of deleterious mutations, which will increase the risk of extinction. However, demographic recovery can still occur after extreme declines, suggesting that natural selection may purge deleterious mutations, even in extremely small populations. The Chatham Island black robin (Petroica traversi) is arguably the most inbred bird species in the world. It avoided imminent extinction in the early 1980s and after a remarkable recovery from a single pair, a second population was established and the two extant populations have evolved in complete isolation since then. Here, we analysed 52 modern and historical genomes to examine the genomic consequences of this extreme bottleneck and the subsequent translocation.

Results: We found evidence for two-fold decline in heterozygosity and three- to four-fold increase in inbreeding in modern genomes. Moreover, there was partial support for temporal reduction in total load for detrimental variation. In contrast, compared to historical genomes, modern genomes showed a significantly higher realised load, reflecting the temporal increase in inbreeding. Furthermore, the translocation induced only small changes in the frequency of deleterious alleles, with the majority of detrimental variation being shared between the two populations.

Conclusion: Our results highlight the dynamics of mutational load in a species that recovered from the brink of extinction, and show rather limited temporal changes in mutational load. We hypothesise that ancestral purging may have been facilitated by population fragmentation and isolation on several islands for thousands of generations and may have already reduced much of the highly deleterious load well before human arrival and introduction of pests to the archipelago. The majority of fixed deleterious variation was shared between the modern populations, but translocation of individuals with low mutational load could possibly mitigate further fixation of high-frequency deleterious variation.

Keywords: Bottleneck; Genomics; Inbreeding; Mutational load; Near-extinction; Translocation.

Fig. 1

Sampling locations and population structure of black robin (Petroica traversi). a Map depicting New Zealand and the Chatham Islands with sampling locations. Circles and triangles represent historical and modern samples, respectively. Although the exact sampling location of historical specimens is sometimes uncertain and incomplete (Table S1), the origin of samples with more accurate sampling information is shown on the map. b Principal Component Analysis (PCA). (c) Admixture plot for K = 2–5. Historical specimens with inconsistent labelling and clustering are shown in the PCA and Admixture plot with asterisks

Fig. 2

Past demography of the black robin (Petroica traversi). a Demographic history of the black robin including historical and modern birds using the PSMC approach assuming a substitution rate of 2.3×10^-9 substitutions/site/generation and a generation time of 2 years. Each curve represents an individual genome. Approximate timing of glacial and interglacial periods are shown in blue and orange, respectively. (b) Demographic history over the past 600 years estimated with GONE. Maung’ Rē and Hokorereoro are grouped into a single population

Fig. 3

Heterozygosity and inbreeding estimates for black robin (Petroica traversi). a Genome-wide autosomal heterozygosity (heterozygous sites / 1,000 bp). Horizontal lines within boxplots depict the mean, bounds of boxes represent the standard deviation and vertical bars represent minima and maxima. b Inbreeding coefficients estimated using ROH (F_ROH). Open bars show the total proportion of the genome in ROH $\ge$ 100 kb and solid bars show proportions in ROH $\ge$ 2 Mb with respective p-values separated by ‘/’. Bars extending from the mean values represent the standard deviation

Fig. 4

Estimates of mutational load based on GERP scores for black robin (Petroica traversi). Individual relative mutational load is based on GERP scores > 5.88 (i.e., derived sites in the top 1% highest GERP score values and considered as highly deleterious). Horizontal lines within boxplots depict the mean, bounds of boxes represent the standard deviation and vertical bars represent minima and maxima

Fig. 5

Estimates of mutational load in coding regions for black robin (Petroica traversi). a Total load. b Realised load. Horizontal lines within boxplots depict the mean, bounds of boxes represent the standard deviation and vertical bars represent minima and maxima

Fig. 6

R_xy of derived alleles for Moderate and High impact variants for black robin (Petroica traversi). R_xy < 1 indicates a relative frequency decrease in population x vs y, for a given variant category (Man. = Maung’ Rē, Hoko. = Hokorereoro, Hist. = Historical). R_xy distributions are based on Jack-knifing across chromosomes