Inbreeding reduces fitness in spatially structured populations of a threatened rattlesnake

Abstract

Small and fragmented populations are at high risk of local extinction, in part because of elevated inbreeding and subsequent inbreeding depression. A major conservation priority is to identify the mechanisms and extent of inbreeding depression in small populations. The eastern massasauga (Sistrurus catenatus) rattlesnake is listed as Federally Threatened in the United States, having experienced significant habitat fragmentation and concomitant population declines over the past 200 years. Here, we use long-term monitoring of two wild populations of eastern massasaugas in Michigan to estimate the extent of inbreeding in each population, identify mechanisms that generate inbreeding, and test for the impact of inbreeding on fitness. Using targeted genomic data and spatial coordinates of capture locations from over 1000 individuals, we find evidence of inbreeding and link inbreeding to spatial kinship structure within populations, possibly driven by limited dispersal. We reconstruct multigenerational pedigrees for each population to measure reproductive output and use long-term capture-recapture data to estimate individual survival (i.e., the two major components of fitness). We find evidence of inbreeding depression in both fitness metrics. The 5% most inbred individuals are 13.5% less likely to have any surviving offspring and have 11.6% lower annual survival compared to all less inbred individuals. By combining genomics and long-term monitoring data, we are able to link the life history of eastern massasaugas to inbreeding and detect relationships between fitness and inbreeding. These insights provide important conservation context for future management and for understanding how spatial structure can generate inbreeding depression even at fine spatial scales.

Keywords: inbreeding; inbreeding depression; pedigrees; population structure.

Fig. 1.

(A) Figure shows the proportion of individuals in Barry and Cass Counties who were assigned no, one, or two parents in pedigree. (B) Histograms show the number of assigned offspring for Barry and Cass Counties. (C) Reconstructed pedigrees of eastern massasaugas in Barry County (Top) and Cass County (Bottom), MI. Genotyped individuals are colored in accordance with F_grm, and nongenotyped “dummy” individuals are outlined in gray. Pedigrees have been simplified for visualization, and some individuals appear multiple times. (D) Histograms show the distribution of inbreeding in Barry and Cass Counties. (E) Family pedigrees show instances of close-kin mating that resulted in offspring with elevated F_grm. Numbers mark the location of focal individuals in both the family and population-wide pedigrees.

Fig. 2.

Top plots show genetic principal component analyses for eastern massasaugas at Barry (A) and Cass (B) County sites. Points are colored according to an individual’s centroid latitude at Barry or longitude at Cass, visualizing the main spatial axis of genetic variation within each site. The maximum distance between individual centroids along a latitudinal axis at Barry County was 1,048 m, and the maximum distance between individual centroids along a longitudinal axis at Cass County was 1,086 m. Bottom plots show distributions of pairwise distances in meters between unrelated pairs without offspring (gray), dam–offspring pairs (orange), sire–offspring pairs (yellow), and parents with offspring in the pedigree (blue) for Barry (C) and Cass (D) County sites. The bar and star on each boxplot represent the median and mean of pairwise distances, respectively. All comparisons between dam–offspring, sire–offspring, and parent pairwise distance distributions and unrelated distance distributions are significant (P < 0.05) using Kolmogorov–Smirnov tests with Bonferroni-corrected P-values.

Fig. 3.

(A) Effect of inbreeding, as measured by F_grm, on annual apparent survival from capture–recapture models of eastern massasaugas from Barry (dashed line) and Cass (solid line) counties. Shaded bands represent 95% CI. Internal ticks mark observed values of F_grm for Barry (black) and Cass (blue) counties. To facilitate site comparisons, variation in snout–vent length (SVL) was controlled for by holding it at 52 cm (moderately sized adults at both sites). (B) Effect of F_grm on the probability of excess zeros in number of pedigree offspring in the zero-inflated model. The solid line represents the predicted probability of offspring. The shaded band represents the 95% CI. Orange points show individuals that did (zeros) and did not (ones) have offspring.