Our lab aims to address fundamental questions in evolutionary and behavioral biology by studying genetic variation in the ways that individuals choose, shape and learn about their environments–especially their social and nutritional environments–and how these processes feed back to influence behavior and fitness.
How do genetic and environmental variation together orchestrate behavior? And, how does behavior evolve when genes determine environments? The dynamic ways that individuals both contribute to, and are shaped by, their environments are widespread; but their implications for behavior, genetics, and evolution are still poorly understood.
To meet these challenges, we use sophisticated behavioral analysis, genetics and genomics, and artificial evolution to understand how genetic and socio-environmental variation together orchestrate behavior across timescales, how these processes influence evolutionary-genetics parameters such as heritability and selection, and ultimately, how they play out over generations to shape evolutionary change
Ecologists and evolutionary biologists typically assume that genotype and environment have independent effects on trait variation, but we also now know that, when genotypes differ in behaviors such as social group choice (Saltz 2011; Saltz 2017; Geiger & Saltz in revision), aggression (Saltz & Foley 2011; Saltz 2013), and habitat choice (Miller et al 2011; Burns & Saltz unpubl.) different genotypes will systematically experience different environments. Such links between individuals’ genotypes and their environments are important because the environment that individuals experience can often influence their later behavior via phenotypic plasticity. Supporting this idea, we found that genotypes develop different levels of aggressiveness when they experience their preferred social environment, compared to an alternate environment (Saltz 2017). We are currently examining how genetic variation in aggressiveness modifies individuals’ early-life experiences and how these experiences carry over to influence subsequent aggressive encounters (Douglas & Saltz, in prep.).
Similarly, we are investigating how genetic and species differences in habitat preferences are modified by learning. Genotypes or species that explore a particular habitat, and/or are particularly sensitive to positive or negative experiences in that habitat, may be more likely to learn than other genotypes (Saltz et al 2017). We are testing these hypotheses on two evolutionary timescales: within populations of D. melanogaster, and across sister species D. simulans and D. sechellia. These species have recently and dramatically diverged in habitat breadth and food preferences; D. sechellia is a specialist on a host, Morinda citrifolia, that is toxic to most insects, including the habitat generalist D. simulans. So far, we have discovered substantial genetic variation in learning in D. melanogaster that genetically co-varies with other types of behavioral plasticity (Saltz et al 2017). Our preliminary findings in the other species suggest that genetic variation in a wide range of behaviors— preferences, exploratory behavior, and responses to stimuli—facilitate or inhibit learning, and that considering these diverse processes can illuminate the reasons that learning evolves.
Quantitative-genetics parameters that govern evolution, like genetic correlations and genotype-by-environment (GxE) interaction, derive not from genetic and developmental processes that cause behavior(s), but from variation in these processes within populations. One of my lab’s long-term goals is to develop theory that allows us to predict this variation a priori (Saltz et al 2017; Saltz et al 2018). Recently, we collaborated to develop statistical models that compare the genetic variation in learning we observed in D. melanogaster (see 1) with patterns expected under Bayesian models of development. Our findings supported the predictions of the model (Stamps et al, in press), suggesting that genetic variation in plasticity (i.e., GxE) may follow predictable “rules.” Our results indicate that Bayesian models of development could enable quantitative predictions about who should learn in which environments, further revealing why populations of many species differ in the magnitude of heritability and GxE (Saltz 2011, Saltz et al 2018).
As with population-genetics parameters, social structures inherently emerge from the behaviors of multiple interacting individuals. For example, we find that genetic differences in aggressiveness influence group size and use of food patches (Saltz & Foley 2011; Foley, Saltz et al 2015), and genotypes tend to form groups with individuals who share similar behaviors (Saltz unpubl). Within groups, the behaviors of a single individual can influence the aggressive and mating behaviors of the entire group (Saltz 2013; Saltz 2017). These findings have implications for the ways we think about and measure social environments (Saltz et al 2016), especially by revealing that there is no such thing as a “standard” social environment.
Currently, we are examining group formation—i.e., which individuals end up in groups together—and how social networks (quantitative patterns of behavioral interactions among individuals) form within groups. Using automated video tracking, we are comparing social network formation in 100 replicate groups of 20 genetically-diverse individuals as they interact in different nutritional environments. So far, we find that (1) genotypes differ in social network parameters; (2) social network parameters differ across nutritional environments; and (3) none of these patterns are explained by variation in overall activity. Coupled with fitness data from the same individuals, these ongoing analyses will reveal some of the causes of variation in social structure and how these contribute to social evolution.
Social structure matters because it shapes the fitness benefits of behaviors and thus the genetic variation present within populations. We are evaluating the mechanisms by which variation in social dynamics produce variable social selection pressures by studying the interplay between social experience and physiological limitations that shape tradeoffs across mating bouts—a process central to sexual selection theory, but surprisingly under-studied (Douglas & Saltz, submitted).
We are also expanding our study of the fitness outcomes of social dynamics to multiple ecological contexts. As noted above, our work on the ecological and genetic determinants of social networks incorporate the effects of nutrition and social network position on mating success and fecundity. In addition, we are examining how group composition influences behavior and resultant outbreaks of a deadly infectious fungus (Keiser et al, 2017, 2018). This latter research goal is a collaboration with former lab postdoc, C. Nick Keiser, now at the University of Florida.