How human behavior can impact the evolution of genetically-mediated behavior in wild non-human species

Abstract

Humans intensely modify the ecosystems we inhabit. Many of the impacts that this behavior can have on other species also sharing these spaces are obvious. A prime example is the devastating current extinction crisis. Yet some populations of non-human, non-domesticated species survive or even appear to thrive in heavily disturbed or human-built habitats. Theoretically, this apparent paradox could be facilitated partly by the evolution of genetically-mediated trait adaptations to the impacts of human behavior. At the least, persistence in strongly modified habitats would provide requisite selection pressures for this process to potentially occur in the future. In fact, we have a growing number of well-characterized examples of morphological trait adaptations to human behavior. However, our knowledge of genetically-mediated behavioral adaptations in similar contexts is less well developed. In this review I set up and discuss several evolutionary scenarios by which human behavior might have impacted the evolution of genetically mediated behavior in non-human, non-domestic species and highlight several approaches that could be used in future studies of this process.

Introduction

Humans profoundly modify and are deeply embedded in the ecosystems we inhabit (Boivin et al. 2016; Worm and Paine 2016), with widespread implications for coexisting non-human, non-domesticated species. Human-mediated landscape changes, harvesting pressures, and the translocation of plants, animals, and pathogens can negatively or positively affect endemic taxa population densities and directly or indirectly lead to extinction (Ceballos et al. 2015).

Human-mediated ecological changes can also affect the evolutionary biology of other species (Hendry et al. 2017), for example through genetically-mediated body mass or feature size reductions in response to selective hunting, fishing, or gathering pressures (e.g., Jachmann et al. 1995; Law and Salick 2005; Pigeon et al. 2016) or through functional morphological adaptations to navigate the human-built environment more efficiently (e.g., Brown and Bomberger Brown 2013; Desrochers 2010; Winchell et al. 2016).

My colleagues and I recently reviewed processes by which aspects of human behavior can effect non-human evolution (Sullivan et al. 2017). Because one of our goals in that review was to connect insights from modern biology to data from the archaeological record, we focused on morphological traits; we concluded that these processes likely extended considerably into prehistory, perhaps to 50,000 years before present or even earlier (Sullivan et al. 2017).

If human behavior has been a long-term ecological driver of the evolution of morphological traits in non-human species, then similar processes have likely affected the evolution of other phenotypes as well. Inspired by my participation in the President’s Symposium of the 2018 International Congress of Neuroethology, in this article I discuss how human behavior may impact the adaptive evolution of genetically-mediated neuroanatomy and behavior in non-human, non-domestic species. Along with highlighting several already-documented examples of this phenomenon, I describe approaches that may be useful in future efforts to identify and study these patterns more broadly.

Example cases of adaptation in response to human behavior

This review is focused on observed and theoretically plausible adaptive responses to human behavior in non-domesticated, non-human species. For the purposes of this review, ‘domestication’ is defined as a process involving the conscious selection of traits in managed breeding populations for their expanded representation in subsequent generations. Later in the review I discuss the utility of potential overlap between patterns of genetic adaptation in non-domesticated species in response to human behavior versus those for now-domesticated taxa (which in some cases, e.g. with dogs, may have begun prior to the onset of the domestication process as it is narrowly defined here).

The null hypothesis for the cause of observed evolution of a genetically-mediated trait should be neutrality (i.e. genetic drift). If adaptive evolution is determined to have occurred, then processes other than responses to human behavior are perhaps more likely to have helped drive this change (e.g. Purdue 1989; Bhat et al. 2020). Even more, there are not yet many well-characterized examples of genetically-mediated behavioral adaptation in non-human, non-domesticated species in response to the impacts of human behavior. Still, a number of promising and likely cases of this phenomenon do exist; these illustrative cases are peppered throughout this review. However, to otherwise help set a more complete stage for the relatively speculative discussion on potential future studies in this area, the current section also includes descriptions of a select few of my favorite morphological trait examples of this general evolutionary process.

Harvesting pressures

Human hunters or gatherers may favor prey body sizes or trophy features at one end of the population trait distribution of the targeted species. If genetic variants underlie at least part of the expression of the trait and if trait-selective harvesting pressures are sufficiently strong, then an evolutionary change may be observed. Famously, at Ram Mountain, Alberta, Canada, the horns of male bighorn sheep (Ovis canadensis) decreased ~30% in size over a 23 year period of intense trophy hunting pressure from 1973–1995 (Coltman et al. 2003). Following a 10-fold decline of sport harvest starting in 1996, horn lengths then began increasing gradually (Pigeon et al. 2016).

The evolutionary effect of human size-selective harvesting pressure on non-human morphological traits is ancient. Coastal shell middens (trash mounds) offer a temporal record of human activity (e.g., fish and shellfish species consumed) and the potential impacts of this behavior on the harvested species (e.g., via inference of prey body sizes). For example, the midden record of Caribbean Panama reveals that fighting conch (Strombus pugilis) have been intensively harvested for at least ~3500 years, with a striking ~40% estimated decrease in the body sizes of sexually mature conch from pre-human intense foraging times to today, likely driven by this process (O’Dea et al. 2014).

Landscape modifications

Non-human species may also evolve in response to human-modified landscapes, including to our built environments. This is a great example: cliff swallows (Petrochelidon pyrrhonota) were studied over a 30 year period in Nebraska, USA, where bridge and culvert construction expanded beginning in the 1980s resulting in increased rates of cliff swallow roadside nesting behavior and vehicle impact risk (Brown and Bomberger Brown 2013). Over this period, wing lengths of road-killed cliff-swallows were significantly longer than those of the overall population, likely because shorter wings aid flight agility and help reduce the risk of vehicle collision. Population wing length decreased ~2% over the study period and the frequency of birds killed by vehicle impacts declined by nearly an order of magnitude despite a concurrent increase in the proportion of sport-utility vehicles, which are otherwise associated with higher impact risk.

The culture of widespread garden bird feeder placement in the United Kingdom has apparently led to the evolution of longer bill lengths to more efficiently exploit this resource in at least two bird species, Central European blackcaps (Sylvia atricapilla) (Rolshausen et al. 2009) and great tits (Parus major) (Bosse et al. 2017). In the great tit study, population genomic data were used to identify genetic regions with signatures of recent positive selection (i.e. increases in allele frequencies significantly more rapidly than that expected under neutral processes), which were enriched for genes with known roles in bill and craniofacial morphology and were cumulatively associated with bill length variation in the study population (Bosse et al. 2017). Moreover, an analysis of radio frequency identification (RFID) bird and feeder data suggested that individuals with longer bills and individuals homozygous for the most prominent long-bill candidate genetic haplotype (encompassing the collagen gene COL4A5) spent significantly more time at bird feeders than their conspecifics (Bosse et al. 2017).

As a potential behavioral adaptation example, Thomsen et al. (2017) described changes in the human biting behavior of the mosquito Anopheles farauti from a four-year longitudinal study that spanned the nationwide distribution of long-lasting insecticidal bednets in Papua New Guinea. Mosquito bite rates in the early evening hours, before people were protected by bednets, increased significantly (Thomsen et al. 2017). There is not yet evidence that the observed behavioral shift is genetically mediated, but this proposition is worth investigating, and with major public health implications.

Indirect effects of non-endemic species translocations

Human-mediated introduction of invasive species can also have evolutionary implications for endemic taxa. For example, the 1935 introduction of toxic cane toads (Bufo marinus) to Australia has likely driven the evolution of both smaller adult head sizes and larger body sizes in two endemic snakes (Pseudechis porphyriacus and Dendrelaphil punctulatus); these traits obviate the consumption of larger/more toxic toads and physiologically reduce the relative toxic dose for any ingestion, respectively (Phillips and Shine 2004).

In at least one case there is evidence for morphological and potentially also genetically-mediated behavioral adaptation in response to an invasive species. Humans facilitated the introduction of venomous fire ants (Solenopsis invicta) from South America into the southern United States by the 1940s, where the ants now prey on endemic fence lizards (Sceloporus undulatus). Lizards in long-invaded sites have significantly longer hindlimbs than those in not-yet-invaded locations (Langkilde 2009), opposite the biogeographic pattern observed from pre-invasion museum collections (Thawley et al. 2019). When combined with a twitching behavior in response to fire ant exposure – also more commonly observed in experiments with lizards from long-invaded sites (with an uncertain but potential genetic component) – longer hindlimbs aid ant removal to benefit survival (Langkilde 2009).

Again, this was only a brief illustration of selected non-human evolutionary changes stemming directly or indirectly from human harvesting pressure, landscape modification, and species translocation behaviors, to help me next discuss several specific hypotheses for how these human behaviors could impart selection pressures on non-human behavior and neuroanatomy. For further morphological examples of these processes and extended discussion please see several recent reviews of this topic both from my group and others (Alberti 2015; Alberti et al. 2017; Hendry et al. 2017; Johnson and Munshi-South 2017; McDonnell and Hahs 2015; Pelletier and Coltman 2018; Perry 2014; Sullivan et al. 2017; Turcotte et al. 2017).

Predicted processes of behavioral adaptation to human impacts

Here I outline a few hypothetical evolutionary scenarios by which human behavior might have impacted the evolution of genetically mediated behavior and/or neuroanatomy in non-human, non-domestic animals. These areas are intended to help highlight interesting areas for targeted future investigations, but this list is not intended to be comprehensive.

I would like to begin with the ‘weed macaque’ concept developed by Richard and colleagues (1989). These authors noted a striking difference between two different sets of species in this genus that cross-cuts phylogenetic relationships. One set includes rhesus macaques (Macaca mulatta), long-tailed macaques (M. fascicularis), and bonnet macaques (M. radiata); these ‘weed’ macaques often live in or nearby human settlements and thrive when foraging on human crops or on flora and fauna that tend to be found at the edges of these habitats. In contrast, the non-’weed’ species, which include the pig-tailed macaque (M. nemestrina) and the lion-tailed macaque (M. silenus), strongly prefer habitats with much less notable signs of human disturbance.

Regardless of whether this variable propensity among macaques towards human-disturbed habitats reflects longstanding (i.e. pre-intense human habitat disturbance) foraging niche differences or a history of behavioral (and potentially morphological) adaptations to human disturbance that has already taken place in some species, this variation could readily lead to further behavioral adaptations in the ‘weed’ species if genetic variation associated with the expression of relevant traits already exists in the populations. Evolutionary comparisons between sets of anthro-philic (‘weed’) and generally anthro-phobic taxa, including with parallels that could likely be drawn for many other taxonomic groups, could be a fruitful direction of research to identify more detailed adaptive hypotheses and to study these processes.

A recent global meta-analysis reported that human disturbance significantly impacts the activity patterns of non-human, non-domesticated species, with a shift towards nocturnality by an average factor of 1.36 among mammals across a broad range of body sizes (Gaynor et al. 2018). While the initial activity pattern shifts themselves likely reflect plastic behavioral responses, such a change might in turn result in altered selection pressures on a suite of other physiological phenotypes and behaviors that could be identified and studied in detail.

Urban evolution is a particularly great case study of non-human adaptation to human disturbance, with insights that could then be translated to help inform studies conducted across less extreme gradients of human disturbance. In city environments, while some non-human species are effectively excluded completely, others seem to thrive. How do these two groups of species differ? Have thriving urban populations genetically adapted to this habitat, behaviorally and/or morphologically?

There are a handful of well-characterized morphological urban evolution examples (e.g., Goiran et al. 2017; Littleford-Colquhoun et al. 2017; Van’t Hof et al. 2016; Winchell et al. 2016), along with several reports of behavioral differences between urban and rural populations that are likely genetically-mediated and thus could have evolved adaptively. In one such study, Miranda and colleagues (2013) hand-raised European blackbirds (Turdus merula) from both urban and nearby rural areas in a common environment. As adults, birds from the urban population were significantly more likely to avoid novel objects placed in a familiar foraging area (Miranda et al. 2013).

Meanwhile, van Dongen and colleagues (2015) surveyed genetic variants in the dopamine receptor gene DRD4 between urban and non-urban populations of black swans (Cygnus atratus) separated by only 30 km. They identified a tentative association between DRD4 genotype and observed wariness behavior (flight initiation distance) variation among the swans, and they reported that DRD4 alleles associated with greater wariness were much less frequent in the urban population, despite otherwise low genetic differentiation between these populations at other markers across the genome (van Dongen et al. 2015).

The artificial structure of urban environments and the noises of the city might also serve as a potential driver of natural selection on intra-specific communication behavior. Two studies identified song structure differences between urban and rural house finches (Haemorhous mexicanus), albeit with co-associated differences in vocalization-influencing bill morphology (Badyaev et al. 2008; Giraudeau et al. 2014). Since bill morphology is also critical for foraging, and food type availability differences between urban and rural environments are expected, the ultimate evolutionary driver(s) of the observed differences (if they do not simply reflect genetic drift) in house finches cannot yet be established. Still, intra-specific communication generally is an excellent opportunity for further studies of the broader process of human impacts on the evolution of genetically-mediated behavior in non-human species sharing our environments.

Practical thoughts for looking forward

If a student was interested in developing a new research project or program in this area, what approaches might they use? My following suggestions are admittedly biased towards genomics-based designs, in part because I find this to be such a valuable tool for helping to test evolutionary hypotheses. Still, different starting points are available. For example, for non-human species with already documented morphological adaptations in response to human behavior, there might regularly be associated behavioral adaptations as well, for example as might be the case for fence lizards in response to invasive fire ants (Langkilde 2009), as described above.

One potentially fruitful approach might be to test whether neuroanatomical traits and genetically-mediated behaviors that have been repeatedly observed in domesticated species have also evolved in parallel in human-associated ‘weed’ taxa, even if to a lesser degree, or in comparisons between (for example) urban vs. non-urban populations within the same species. As a starting point, a suite of possibly linked anatomical and behavioral traits in domesticated taxa has been tentatively linked to common underlying changes in neural crest cells (Sanchez-Villagra et al. 2016; Wilkins et al. 2014). As we continue to obtain more detailed understandings of the convergent genetic and molecular processes observed among domesticated taxa (e.g., Alberto et al. 2018; Carneiro et al. 2014; Kukekova et al. 2018; Librado et al. 2017; Pendleton et al. 2018; Wang et al. 2018), applying this knowledge to the potential study of adaptations in anthro-philic non-human populations or species will become increasingly feasible.

Genomic scans for ‘signatures’ of past positive natural selection, which now can readily be conducted in practically any non-human species, can also ultimately help researchers ‘reverse engineer’ suites of potential adaptations to human behavior. Effectively, genomic-scale sequencing or genotype data from one (e.g., an urban population of a particular species) or multiple populations (e.g. comparing urban and rural populations) are analyzed to identify particular genetic regions in which variants within have increased in frequency significantly more rapidly than can be readily explained by genetic drift processes (or that are merely outliers in the distribution of these differentiation statistics). Then, one can test whether phenotypic and/or behavioral variation is associated with these regions. Perhaps in some cases, relevant trait hypotheses might even only be developed following review of functional genetic knowledge of the identified candidate regions.

The above approach was effectively used by Bosse et al. (2017) in their study of bill length adaptation in United Kingdom great tits, discussed above. Yet there are other regions of the genome also containing strong signatures of past positive natural selection, beyond those that they associated with bill length variation, some of which might underlie behavioral or other morphological traits that also reflect adaptation to provisioned feeder foraging. Elsewhere, Harris and Munshi-South (2017) performed a genome-wide scan for signatures of positive natural selection in three urban and three rural populations of white-footed mice from within and outside New York City. Among the candidate regions identified, they noted enrichments for genes involved in metabolic processes, possibly reflecting diet-related adaptations in the urban mice (Harris and Munshi-South 2017). Additional insights may still emerge from detailed investigations of other candidate loci uncovered from their analysis.

I expect that this type of ‘reverse engineering’ approach, along with increasingly detailed associated functional and fitness-related investigations, will become more and more commonly applied in the coming years. In many cases it might also be possible to then incorporate historic or ancient DNA data from museum or archaeological contexts into this analytical framework (Marciniak and Perry 2017), in order to powerfully track the identified phenotype- or behavior-associated alleles over time and across space. I look forward to seeing the progress in this field!