Research ArticleEVOLUTIONARY BIOLOGY

Genomic signatures of extensive inbreeding in Isle Royale wolves, a population on the threshold of extinction

See allHide authors and affiliations

Science Advances  29 May 2019:
Vol. 5, no. 5, eaau0757
DOI: 10.1126/sciadv.aau0757
  • Fig. 1 Isle Royale wolf pedigree and geographic origins of genomes in this study.

    (A) Pedigree of Isle Royale wolves sequenced in this study (numbered individuals), adapted from Hedrick et al. (25). Circles represent females, and squares represent males. Relationships were inferred from genotypes at 18 microsatellite loci. Shaded individuals were examined for the presence of vertebral abnormalities for this study (see Table 1). The ancestries of F25, F55, M61, and F67 are unknown, but F67 is known to be the mother of F65 and M175. M was a wolf that migrated from the mainland in 1997 (24). The “A” individuals were the last two wolves alive on Isle Royale in 2018. (B) Map showing approximate origins of all individuals analyzed in this study. BI, Baffin Island; EI, Ellesmere Island; NU, Nunavut; VI, Victoria Island. See table S1 for further sample information.

  • Fig. 2 Distributions of heterozygosity across the genome under different demographic histories.

    In each panel: (left) Example barplots showing per-site heterozygosity in nonoverlapping 1-Mb windows across the autosomal genome; (right) histograms of per-window heterozygosity. (A and B) The Xinjiang and Minnesota wolves represent large outbred populations. (C and D) The Tibetan and Ethiopian wolves represent small isolated populations. (E and F) The Mexican and Isle Royale wolves represent populations with recent inbreeding. See fig. S2 for plots of all individuals. Het., heterozygotes.

  • Fig. 3 ROH and the correspondence with genome-wide heterozygosity and FPED.

    (A) Left: Per-site autosomal heterozygosity across the autosomal genome. Samples are ordered by decreasing heterozygosity from top to bottom. Right: Summed lengths of short (0.1 Mb ≤ ROH < 1 Mb), medium (1 Mb ≤ ROH < 10 Mb), and long (10 Mb ≤ ROH < 100 Mb) ROH per individual. (B) FROH is the proportion of the genome contained within ROH of ≥100 kb. FPED values were calculated from the pedigree of Hedrick et al. (25) (Fig. 1A). The gray dashed line shows the diagonal y = x. (C) FROH in Isle Royale wolves over time (by birth year; see table S1) reflects the effects of increasing inbreeding over time before and after the arrival of the migrant individual in 1997 (gray dashed line; see Fig. 1A). Gray bars indicate the census population size on Isle Royale. (B and C) Open circles denote individuals not related to the migrant, and closed circles denote individuals descended from the migrant (according to the pedigree).

  • Fig. 4 Genotype and allele frequencies in Isle Royale versus mainland Minnesota wolves.

    (A) Inbred Isle Royale wolves contain significantly fewer heterozygotes and (B) significantly more homozygotes than Minnesota wolves, for both damaging and benign SNPs. (C) The total number of derived alleles per individual is unaffected by recent inbreeding. ***P < 0.001. NS, not significant. (D to G) Two-dimensional allele frequency spectra showing the correlation in derived allele frequencies (AF) between outbred North American wolves (Quebec, Yellowstone, and Canadian Arctic excluding Ellesmere Island) and Isle Royale or Minnesota wolves for variants present within Isle Royale or Minnesota wolves. Derived nonsynonymous variants occur at low frequencies due to negative selection in large outbred populations, (D and F) but nonsynonymous variants segregating in Isle Royale wolves occur at much higher frequencies due to drift and high relatedness among individuals (E and G). This effect is apparent for putatively damaging and benign variants. All sites were randomly down-sampled to include exactly five individuals in each group. Color indicates the density of points (see legends). The dashed line represents the diagonal y = x, and the solid line represents the regression line (see table S3). (D) 4387 sites, (E) 3109 sites, (F) 47,784 sites, and (G) 35,979 sites.

  • Fig. 5 Model and results from simulations of deleterious variation.

    (A) Demographic model used to simulate the expected number and age of mutations in a large population (North America, 17,350 individuals) versus a small population (Tibet, 2500 individuals). Both populations split from a large ancestral population (45,000 individuals) 12,500 years before the present (3-year generation time). Population sizes and split time from Fan et al. (31). Model not drawn to scale. (B and C) Results from simulations were grouped according to dominance and selection coefficients. Additive, h = 0.5; recessive, h = 0; strongly deleterious, Nes > 100; moderately deleterious, 100 ≥ Nes > 10; weakly deleterious, 10 ≥ Nes > 0; neutral, Nes = 0. (B) The average number of homozygous derived genotypes and total derived alleles per individual (ind.). (C) The average ages of segregating mutations in each population.

  • Table 1 Vertebral phenotypes of Isle Royale wolves.

    Six individuals were examined for vertebral defects, and their phenotypes are listed. The skeletal remains of the other Isle Royale wolves sequenced in this study (F25, F55, M62, F67, and M141) were not recovered and were therefore unavailable for examination. LSTV, lumbosacral transitional vertebra; SCTV, sacrococcygeal transitional vertebra; TLTV, thoracolumbar transitional vertebra; N/A, not applicable.

    IDVertebrae
    examined
    PhenotypeOther
    observations
    M61Atlas-SacrumNo
    abnormalities
    N/A
    F65Atlas-T12,
    L1-Co3
    Minor
    phenotypic
    LSTV
    One thoracic
    vertebra
    missing from
    collection
    F75Atlas-Co16LSTV, TLTV,
    SCTV, two extra
    ribs,
    osteophytes
    All of F75 pups
    had an extra
    presacral
    vertebra, seven
    of the eight
    pups had extra
    ribs
    M152Atlas-SacrumTLTV, SCTV,
    extra rib,
    osteophytes
    N/A
    M175Atlas-Co9LSTV, TLTV,
    SCTV,
    asymmetry at
    Co2,
    osteophytes
    Some ribs are
    missing from
    collection
    F189Atlas-Co2TLTV, extra
    vertebra, SCTV,
    cervical
    intrasegmental
    transitional and
    asymmetry
    Some ribs are
    missing from
    collection. The
    cervical
    malformation is
    the same type
    observed in
    specimen 3529
    [see fig. 3 of
    (8)].

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/5/5/eaau0757/DC1

    Fig. S1. Cladogram of genome sequences in this study.

    Fig. S2. Genome-wide distributions of heterozygosity in all individuals.

    Fig. S3. ROH sharing among Isle Royale wolves across the autosomal genome.

    Fig. S4. Number of ROH per individual for different ROH length categories.

    Fig. S5. Proportion of derived homozygotes in ROH as a function of the proportion of the genome within ROH.

    Fig. S6. Candidate genes underlying Isle Royale phenotypes associated with HPO terms related to skeletal anatomy.

    Fig. S7. Comparison of ROH identified in VCFtools and PLINK.

    Table S1. Sample information for sequences included in this study.

    Table S2. Notes and information from the literature about the demographic history of populations included in this study.

    Table S3. Parameters of linear regression models for two-dimensional allele frequency spectra.

    References (6776)

  • Supplementary Materials

    This PDF file includes:

    • Fig. S1. Cladogram of genome sequences in this study.
    • Fig. S2. Genome-wide distributions of heterozygosity in all individuals.
    • Fig. S3. ROH sharing among Isle Royale wolves across the autosomal genome.
    • Fig. S4. Number of ROH per individual for different ROH length categories.
    • Fig. S5. Proportion of derived homozygotes in ROH as a function of the proportion of the genome within ROH.
    • Fig. S6. Candidate genes underlying Isle Royale phenotypes associated with HPO terms related to skeletal anatomy.
    • Fig. S7. Comparison of ROH identified in VCFtools and PLINK.
    • Table S1. Sample information for sequences included in this study.
    • Table S2. Notes and information from the literature about the demographic history of populations included in this study.
    • Table S3. Parameters of linear regression models for two-dimensional allele frequency spectra.
    • References (6776)

    Download PDF

    Files in this Data Supplement:

Stay Connected to Science Advances

Navigate This Article