Genetic Rescue – Fitness and genomic consequences

As a PhD student studying the effects of genetic diversity overall and immunogenetic diversity specifically on survival and reproductive success in an endangered primate in captive and wild populations, I thought a lot about the potential effects of inbreeding and outbreeding depression. I read literally 100s of papers on the topic. Inbreeding depression describes the negative fitness effects that can occur in small populations when relatives breed with each other for multiple generations, thus genetic diversity is lost through genetic drift and negative alleles are expressed. Outbreeding depression, by contrast, is the negative consequence of breeding two genetically distinct populations leading to a loss of local adaptation. Concerns about outbreeding depression are one of the major theoretical limitations to re-introductions and attempts at ‘genetic rescues’ when small populations and/or endangered species might be suffering from inbreeding depression. For the most part, however, evidence of outbreeding depression has mostly been limited to plants and captive or laboratory studies. Earlier this year, however, Dr. Sarah Fitzpatrick and her co-authors documented an extremely cool example of genetic rescue in populations of wild Trinidadian guppies, contradicting the hypothesis about the potential for maladaptive gene flow in population introductions (Fitzpatrick et al. 2020).

Dr. Sarah Fitzpatrick, lead author.
Photo Credit:

Trinidadian guppies. Photo Credit:

After repeatedly sampling two isolated guppy ‘recipient’ populations (Figure 1A, dark blue circles, N < 100 individuals per population) in the Caigual and Taylor rivers in Trinidad, the authors introduced populations of guppies upstream (dashed red circles) of these recipient populations, in previously guppy-free areas. These trans-located guppies, from downstream populations (solid red circles), occasionally (or frequently!) migrated downstream into the recipient populations located either ~5m or ~700m from the introduction location. For ~8-10 guppy generations after the trans-location, the recipient populations have been monitored with mark-recapture to assess population size as well as individual overall genetic diversity, hybrid ancestry, lifespan, and reproductive success. Following the onset of immigration and subsequent gene flow, both recipient populations experienced nearly a 10-fold increase in population size, from less than 100 individuals to an estimated 1,000 individuals each (Figure 1B). Based on the hybrid index, which ranges from 0 to 1 based on the amount of native or immigrant ancestry of an individual respectively, of the generations, it’s clear that 10 generations after the first wave of immigration, the population consists almost entirely of admixed individuals (Figure 1C).

Figure 1 – Gene Flow Manipulation Experiments in Trinidad
(A) Map of the Guanapo River drainage. In 2009, guppies were translocated from a downstream high-predation locality (red) into two headwater sites (dashed red) that were upstream of native recipient populations in low-predation environments (dark blue). Unidirectional, downstream gene flow began shortly after the introductions, indicated by black arrows.
(B) Census sizes in Caigual (solid) and Taylor (dashed) following the onset of gene flow from the upstream introduction sites. Gray box indicates the time span in which all captured individuals were genotyped at 12 microsatellite loci.
(C) Temporal patterns of continuous hybrid index assignments throughout the first 17 months of the study (∼four to six guppy generations). Individuals from recipient populations prior to gene flow had a hybrid index = 0, and pure immigrant individuals had a hybrid index = 1. Hybrid indices were assigned using data from 12 microsatellite loci. Red arrows indicate the onset of gene flow.

Contradicting the predictions of outbreeding depression, individuals with intermediate to high (0.5-0.75) hybrid indices had the highest longevity and reproductive success in both locations and across sexes (Figure 2). Interestingly, although hybrids and pure immigrants had similar levels of genetic heterozygosity, hybrids had higher fitness, suggesting that increased genomic diversity alone does not explain the increased fitness and pointing towards a potential maintenance of locally adapted alleles.

Figure 2 – Relationships between Hybrid Index and Fitness
Fitness metrics (longevity and total lifetime reproductive success [LRS]) varied quadratically with hybrid index (0, pure recipient genotype; 1, pure immigrant genotype). Maxima of the quadratic functions are indicated by vertical dashed lines/diamonds; uncertainty in their positions is indicated by (horizontal) 95% confidence bars. Shading around regression lines displays approximate 95% confidence bands obtained through simulation.
(A and B) Longevity differed between males (red) and females (blue). Generally, females lived longer than males, and fish in (A) Caigual lived longer than those in (B) Taylor. In Taylor, male and female longevity had quadratic relationships with hybrid index that differed in magnitude but peaked at similar parameter estimates; this differed by sex in Caigual (A versus B).
(C and D) LRS varied quadratically with hybrid index, and this trend did not differ between males and females. Individuals from Taylor generally had lower LRS than Caigual (C versus D) and were more likely to not reproduce at all, especially those with recipient genotypes (hybrid indices near zero).

Pre-introduction, 95% and 96% of >12,000 genotyped SNPs were monomorphic in the Caigual and Taylor populations respectively and average nucleotide diversity was 0.01 in both populations (Figure 4b). 8-10 generations later, only 22 and 24% of SNPs are monomorphic and nucleotide diversity has increased to 0.21 and 0.22. Genome-wide average Fst between source and recipient populations also decreased from 0.29-0.31 to 0.01.

To determine if gene flow swamped locally adaptive variants, the authors identified 146 loci with allele frequencies in the pre-immigrant recipient populations that might indicate candidacy for locally adapted alleles. Post-immigration, although overall genome homogenization increased between immigrant and recipient populations, the authors found evidence for selective maintenance of some of the candidate alleles in the recipient populations in the form of an excess of pre-immigrant ancestry at these loci (Fig 4). Unfortunately, none of these candidate loci matched previously identified loci under selection nor were any gene ontology terms enriched, but they provide interesting potential targets for future investigation.

Figure 4 – Genomic Consequences of Gene Flow
New gene flow caused overall genomic homogenization, but candidate adaptive alleles were maintained at higher than expected frequencies.
(A) PCA plot showing overall population differentiation based on polymorphic SNP loci from the RAD-seq data.
(B) Comparison of nucleotide diversity patterns along linkage group two among pre-gene flow (dark blue) and post-gene flow (light blue) Caigual (solid) and Taylor (dashed) populations and the introduction source (red). Similar patterns were found across all 23 linkage groups.
(C) Distributions of ancestry-polarized deviations in candidate loci versus frequency-matched non-candidates for both populations. In each stream, the allele frequencies of the candidate loci were significantly closer to the headwater ancestral frequency compared to a set of frequency-matched non-candidates.

This study documents the phenomenon of genetic rescue in two multi-generational wild populations, showing that contrary to expectations, gene flow does not necessarily swam local adaptation, and actually can significantly increase fitness in the form of longevity and reproductive success, subsequently substantially increasing population size. Further, at laest some locally adapted loci appear to have been maintained in both Caigual and Taylor, despite a 10-fold difference in the number of immigrants to each population, suggesting a range of gene flow rates might still allow the maintenance of local adaptation, with extremely important and interesting implications for future conservation-based introduction efforts.


Fitzpatrick, S.W., G.S. Bradburd, C.K. Kremer, P.E. Salerno, L.M. Angeloni, W.C. Funk (2020) Genomic and fitness consequences of genetic rescue in wild populations. Current Biology 30: 517-522.e5.

This entry was posted in conservation, genomics, hybridization and tagged , , , , . Bookmark the permalink.