Multi-generational benefits of genetic rescue

Abstract

Genetic rescue-an increase in population fitness following the introduction of new alleles-has been proven to ameliorate inbreeding depression in small, isolated populations, yet is rarely applied as a conservation tool. A lingering question regarding genetic rescue in wildlife conservation is how long beneficial effects persist in admixed populations. Using data collected over 40 years from 1192 endangered Florida panthers (Puma concolor coryi) across nine generations, we show that the experimental genetic rescue implemented in 1995-via the release of eight female pumas from Texas-alleviated morphological, genetic, and demographic correlates of inbreeding depression, subsequently preventing extirpation of the population. We present unequivocal evidence, for the first time in any terrestrial vertebrate, that genetic and phenotypic benefits of genetic rescue remain in this population after five generations of admixture, which helped increase panther abundance (> fivefold) and genetic effective population size (> 20-fold). Additionally, even with extensive admixture, microsatellite allele frequencies in the population continue to support the distinctness of Florida panthers from other North American puma populations, including Texas. Although threats including habitat loss, human-wildlife conflict, and infectious diseases are challenges to many imperiled populations, our results suggest genetic rescue can serve as an effective, multi-generational tool for conservation of small, isolated populations facing extinction from inbreeding.

Keywords: Puma concolor; Endangered species conservation; Fitness; Florida panther; Genetic rescue; Inbreeding depression.

Figure 1

Ancestry of adult and subadult Florida panthers (n = 547) sampled 1981–2020 in Florida, USA. Ancestry was determined via a clustering analysis of microsatellite genotype data using the program STRUCTURE with K = 2 clusters (canonical and admixed). Cohorts include Florida panthers born during the pre-genetic rescue (Pre1 and Pre2) and post-genetic rescue (Post1–3) periods (see “Methods”). Panthers were designated as canonical if their canonical ancestry q-value was ≥ 90%. Mean q-values for each ancestry category are presented to demonstrate changes in the composition of population ancestry across generations of panthers pre- and post-genetic rescue. The SE values are reflective of the combined sample within each cohort, since q-values presented are proportions of that sample. Values above each bar represent the number of panthers assigned to that ancestral category in the pre- or post-genetic rescue cohort.

Figure 2

Proportion of Florida panthers sampled from 1981 to 2021 in Florida, USA, that exhibited correlates of inbreeding depression for morphological and physiological traits. Comparisons are made via panthers born in cohorts pre- and post-genetic rescue (a) and between defined ancestral categories (b). The percent abnormal sperm data are from Penfold et al.. Statistical analyses are detailed in SI Appendix 3. Different letters in Frame (a) indicate significant differences (P < 0.05; binomial regression analyses with post-hoc comparisons) between cohorts of panthers. The Post3 cohort includes data from 2021 for these correlates of inbreeding. We used two proportion Z-tests to determine significant comparisons (*P < 0.05) between ancestral categories on Frame (b) (SI Appendix 3).  ASD = Atrial Septal Defect and Crypt = Cryptorchidism.

Figure 3

Metrics of genetic variation calculated using genotype data from 16 microsatellite loci in cohorts of Florida panthers pre- and post-genetic rescue (Frames a and b) and ancestral groups (Frames c and d) sampled from 1981 to 2020 in Florida, USA. Values are means and standard errors. Data from Western populations of puma are presented for comparative purposes. Metrics include number of alleles (N_a), number of effective alleles (N_effect), observed and expected heterozygosity (H_o, H_e), allelic richness (A_r), and individual heterozygosity (H_ind). Statistical analyses methods and results comparing these groups are presented in SI Appendix 4.

Figure 4

Assessment of the heterozygosity-fitness correlations between individual heterozygosity (H_ind) and the probability of presence of kinked tails, cowlicks, and cryptorchidism in Florida panthers. Frames (a–c) depict comparisons of panthers categorized as canonical (solid line) versus admixed (dashed line). Frames (d–f) represent comparisons between cohorts of panthers from the pre- and post-genetic rescue periods. Shaded areas represent 95% confidence intervals.

Figure 5

Estimates of the range-wide population size of adult and subadult Florida panthers from 1981 to 2020 using: (1) the 95% lower confidence interval of the motor vehicle mortality (MVM) model-averaged abundance estimate via the method of McClintock et al. plotted in blue for the period 2000–2018 (see SI Appendix 6); (2) the minimum population count (MPC) index of McBride et al. and McBride and McBride plotted in orange, for the period 1981–2015; (3) and the integrated population model (IPM) of Merriell in plotted in grey, for the period 1982–2020. The year in which genetic rescue was initiated (1995) coincides with the subsequent increase in the population size that was documented by all three metrics.

Figure 6

Principal coordinate analysis (PCoA) based on Nei’s genetic distance matrices between (a) canonical and admixed Florida panthers and Western puma populations, and (b) Pre1, Pre2, Post1, Post2, and Post3 Florida panthers and Western puma populations. Ovals highlight the distinction of the cluster of Florida panther samples from the Western puma populations, including Texas. The arrows point to two uncollared Florida panther samples (UCFP010 and UCFP011; circles enlarged for clarity) collected in Palm Beach County Florida in 1983 and 1984, respectively. Both were suspected of being captive pumas that had been released or escaped into the wild. Those suspicions are further corroborated by our genetic analyses and the clustering of these two samples with non-Florida panthers. Variance explained by each PCoA axis is displayed in axis titles.