During this fellowship, I performed two experiments using zebrafish (Danio rerio) as model species. The first experiment covered the first aims of the project by comparing phenotypes and genotypes of fish within fish population as well as their potential to evolve and by examining the density-dependent effects on phenotypic expression and genetic basis. The second experiment covered the second aims of the project by investigating the phenotypic response of the fish population after one generation of harvest-associated selection and/or density reduction.
During the first experiment, I used adult zebrafish derived from semi-domestic parents to breed in a controlled factorial design (4 groups of 3 males crossed reciprocally to 3 females) and produce 36 families. After hatching, each family was divided into two densities: normal density (based on standard protocols) and low density (half normal density). After 6 months, the fish were individually tagged. Over the next 3 months, I collected a series of phenotypic data on 400 fish per density. I measured their mass and length and estimated their growth. I measured their aerobic metabolism (through respirometry analysis) and estimated their maximum and standard metabolic rate, as well as aerobic scope, and measured their swimming performance. In addition, I analyzed each fish behavior using a series of behavioral assays. I measured their activity level, risk taking and curiosity, aggressiveness and sociability. I was then able to apply quantitative genetics models to estimate heritability of each trait, as well as phenotypic and genetic correlation among the traits, in each density environment.
After the phenotypic data collection, I run a series of simulated fisheries capture over 6 weeks to mimic commercial fisheries. The simulations were carried out using a small-scale trawl net in a swimming flume. Each week, I conducted a series of simulations on the fish population in each density to identity the most vulnerable fish to capture. Each week, the 20% most vulnerable was removed from the subsequent simulations. At the end of the 6 week, I was able to identify the 20% of the population that was the least vulnerable to capture. I then took DNA samples from the fish the most and the least vulnerable in both densities to screen for differential selection of genetic variants. I then extracted the DNA and prepared libraries for sequencing.
For the second experiment, I used the fish from the 20% of the population the least vulnerable (the fish that would remain in a fishery under heavy pressure) and control fish (fish that would not have experience fishing) from each density to create a first generation of harvest-associated selection and 20 families. After hatching, each family was divided into the two densities. The fish were then tagged at 6 months and I collected the same phenotypic data as in the parental generation (growth, physiology, behavior). In addition, I took muscle samples from fish of each family and density to investigate the epigenomic modifications (DNA methylation) that could be involved in response to fishing pressure. I then extracted the DNA and prepared libraries for sequencing.
Analyses of the behavioral and sequencing data are still ongoing in both generations. However, preliminary results on physiology revealed that within a generation, fishing can indeed induce selection on growth but also on swimming performance and aerobic metabolism and that the traits under selection are also having significant heritability. This result suggests that fishing can indeed induce a long-term evolution of the physiological traits. In addition, even if the population density doesn’t modify the direction of the selection, the population density can alter the heritability of the traits under selection, shifting the evolutionary trajectory of the population. Across generations, the preliminary results revealed that parental experience in terms of both harvest and population density can influence offspring metabolism. This result suggests that transgenerational responses can also occur immediately after one generation of harvest-associated selection and density reduction, so prior to genetic evolution. Adding information on individual behavior and DNA modification (both genomic and epigenomic) will allow a more complete picture of the outcomes of FIE.
The preliminary results of this project were presented in several meetings including two international conferences (The meeting of the Society of Experimental Biology 2018 and Evolution Montpellier 2018) and will be presented in two other conferences this summer (The meeting of the Society of Experimental Biology 2019 and the 2019 Congress of the European Society for Evolutionary Biology) which enable me to reach a broad audience of both physiologists and evolutionary biologists.
