Environmental variation predicts patterns of phenotypic and genomic variation in an African tropical forest frog
Running title: Mapping environmentally-associated variation in a tropical forest frog
Courtney A. Miller1*, Geraud Canis Tasse Taboue2,3, Eric B. Fokam2, Katy Morgan1, Ying Zhen4, Ryan J. Harrigan4, Vinh Le Underwood4, Kristen Ruegg4, Paul R. Sesink Clee5, Stephan Ntie6, Patrick Mickala6, Jean Francois Mboumba6, Trevon Fuller4, Breda M. Zimkus7, Thomas B. Smith4,8, Nicola M. Anthony1
1. Department of Biological Sciences, University of New Orleans, New Orleans, LA, USA 2. Department of Zoology and Animal Physiology, University of Buea, Buea, Cameroon 3. Institute of Geological and Mining Research, Yaoundé, Cameroon
4. Institute of Environment and Sustainability, University of California, Los Angeles, CA, USA
5. Department of Biology, Drexel University, Philadelphia, Pennsylvania, USA
6. Department of Biology, University of Science and Technology of Masuku, Franceville, Gabon
7. Museum of Comparative Zoology, Harvard University, Cambridge, MA, USA
8. Department of Ecology and Evolutionary Biology, University of California, Los Angeles, CA, USA
ABSTRACT Central African rainforests are predicted to be disproportionately affected by future climate change. How species will cope with these changes is unclear, but rapid environmental changes will likely impose strong selection pressures. Here we examined environmental drivers of phenotypic and genomic variation in the central African puddle frog (Phrynobatrachus auritus ) to identify areas of elevated environmentally-associated turnover where populations may have the greatest capacity to adapt. We also compared current and future climate models to pinpoint areas of high genomic vulnerability where allele frequencies will have to shift the most in order to keep pace with future climate change. Analyses of body size, relative leg length, and head shape suggest that seasonal aspects of temperature and precipitation significantly influence phenotypic variation, whereas geographic distance and precipitation seasonality are the most important drivers of SNP allele frequency variation. However, neither landscape barriers nor the effects of past Pleistocene refugia had any influence on genomic differentiation. Most phenotypic and genomic differentiation coincided with key ecological gradients across the forest-savanna ecotone, montane areas and a coastal to interior rainfall gradient. Areas of greatest vulnerability were found in the lower Sanaga basin and southeastern region of Cameroon. In contrast with past conservation efforts that have focused on hotspots of species richness or endemism, our findings highlight the importance of preserving environmentally heterogeneous landscapes to preserve putatively adaptive variation and ongoing evolutionary processes in the face of climate change.
KEYWORDS: central Africa, amphibians, RAD-seq, environmental gradients, genomic vulnerability, climate change
1. INTRODUCTION
The tropical forests of the Congo Basin and Gulf of Guinea represent one of the most biologically diverse regions in the world. This region ranks third in plant, mammal, bird, and amphibian species richness after the Amazon and New Guinea (Mittermeier et al., 2003). With respect to amphibians, the Cameroonian highlands are recognized as one of the world’s most important biodiversity hotspots (Amiet, 2008; Gvoždík et al., 2020; Herrmann et al., 2005). Several hypotheses have been advanced to explain the high biodiversity in this region. Previous phylogeographic studies have shown that past Pleistocene refugia shaped population structure in several central African rainforest species, providing support for their role as potential engines of diversification (Anthony et al., 2007; Born et al., 2011; Murienne et al., 2013; Nicolas et al., 2011; Plana, 2004; Quérouil et al., 2003). Alternatively, the riverine barrier hypothesis has argued that rivers could have led to the isolation and diversification of tropical forest species (Colyn et al., 1991; Wallace, 1854). Support for this hypothesis has been found in primates (Anthony et al., 2007; Mitchell et al., 2015; Telfer et al., 2003), birds (Aleixo, 2004) and rodents (Nicolas et al., 2011). In contrast, far less attention has been paid to the impact of historical barriers to gene flow or the effects of environmental variation (i.e. precipitation, temperature, vegetation density, etc.) on patterns of amphibian diversification in this region.
Several general hypotheses have been advanced to explain body size variation in amphibians. First, intra-specific amphibian body size has been shown to increase with decreasing temperatures, consistent with the heat balance hypothesis which predicts a negative relationship between body size and measures of energy availability (Boaratti & Da Silva, 2015; Ficetola et al., 2010; Rivas et al., 2018), although this has been found to not always be the case (Goldberg et al., 2018; Guo et al., 2019). Second, several studies have shown that larger bodied animals are found in drier areas (Goldberg et al., 2018; Guo et al., 2019; Olalla-Tárraga et al., 2009) providing support for the water availability hypothesis which predicts that larger bodied animals are less prone to desiccation due to reduced surface-to-volume ratio. In contrast, the converse water availability hypothesis argues that body size is positively related to water availability such that larger animals will be found in areas with greater precipitation (Zug et al., 2001). Finally, the primary productivity hypothesis posits that larger body sizes will be found in areas with greater resource availability (Olalla-Tárraga & Rodríguez, 2007). In support of this hypothesis, Ficetola et al. (2010) found that body size increased in areas of higher primary productivity suggesting that less productive areas may have insufficient food supplies to support larger body sizes. However, aside from a study by (Charles et al. (2018) on the African leaf folding frogAfrixalus paradorsalis , few studies have addressed morphological variation in central African tropical frogs.
Environmental variation can act as a strong agent of diversifying selection, particularly in areas of high environmental heterogeneity (Endler, 1973), such as that observed across ecotones (Freedman et al., 2010; Smith et al., 1997; Termignoni-García et al., 2017) or across different levels of elevation (Thomassen et al., 2011). In central Africa, environmental variation has been shown to explain patterns of genetic differentiation in olive sunbirds (Smith et al., 2011), little greenbuls (Smith et al., 1997; Zhen et al., 2017), skinks (Freedman et al., 2010), chimpanzees (Mitchell et al., 2015), forest antelope (Ntie et al., 2017), soft-furred mice (Morgan et al., 2020), and reed frogs (Bell et al., 2017). These heterogeneous environments may capture ecological and evolutionary processes that are fundamental to maintaining and generating biological diversity (Moritz et al., 2000).
One major challenge is being able to effectively partition the effects of isolation by environment (IBE) from other potential drivers of population differentiation, namely: isolation by distance (IBD) and isolation by resistance due to landscape barriers (IBB). Advances in landscape genomics can be used to simultaneously assess the relative importance of competing ecological and historical drivers on genomic differentiation (Manthey & Moyle, 2015; Termignoni-García et al., 2017). Specifically, Fitzpatrick and Keller (2015) have shown that Generalized Dissimilarity Modelling (GDM; Ferrier et al., 2007) and Gradient Forests (GF; Ellis et al., 2012) can be powerful tools for analyzing gene–environment associations at the landscape level. Under a model of IBE, genetic differentiation increases with environmental differences, independent of geographic distance (Shafer & Wolf, 2013; Wang & Bradburd, 2014). In contrast, under a model of IBD, genetic differentiation is predicted to increase as a function of geographic distance whereas IBB is driven by landscape barriers to animal or plant dispersal (Balkenhol et al., 2017; Cushman & Schwartz, 2006). Resistance distances due to barriers between populations can be based on landscape features that may influence gene flow, such as habitat type (Tucker et al., 2017), physical barriers such as rivers (Mitchell et al., 2015) or the effects of past refugia (Ntie et al., 2017), which hereafter will be referred to as isolation by Pleistocene refugia (IBP). In many of these cases, circuit theory is used to incorporate IBE and IBB into models of population connectivity and identify which variables are the most important predictors of gene flow (McRae & Beier, 2007).
Central Africa faces a variety of threats from human activities and is especially vulnerable to climate change (Abernethy et al., 2013; James et al., 2013; Laporte et al., 2007; Oates et al., 2004). Temperatures are expected to rise along with potential shifts in rainfall patterns, including more intense dry seasons that could result in forest retreat (James et al., 2013). Species in this region, if they are to survive, would therefore be forced to respond to climate change either through dispersal, evolutionary adaptation or phenotypic plasticity (Davis et al., 2005; Holt, 1990). Given the threat that climate change poses to many species, there is now an increasing need to identify current and historical drivers of evolutionary diversification and recognize key areas for future conservation where species capacity to adapt is greatest (Anthony et al., 2015). Mapping landscape-level predictions of environmentally-associated genomic variation under both current and projected future environments can shed light on both the ability of populations to persist in their current state as well as their future capacity to respond to change through evolutionary adaptation (Gunderson, 2000; Sgro et al., 2010; Thrush et al., 2009). In this regard, the “genomic vulnerability” (Bay et al., 2018; Ruegg et al., 2018) has been used as a measure of the degree of “mismatch”, or “offset”, between current and future projections of environmentally-associated genomic variation and can be used as a proxy for population vulnerability to environmental change (Ruegg et al., 2018). Genomic vulnerability estimates how strongly allele frequencies would have to change to keep track with the environmental changes predicted to occur at a certain location. Thus, the locations with the greatest vulnerability are those with the strongest predicted changes in allele frequencies, since these are the places where adaptation will have to keep up with environmental change the most.
In the present study, we used a combination of statistical methods to determine the potential drivers of phenotypic and genomic diversification in the widespread puddle frog, Phrynobatrachus auritus across its range in the west Central African countries of Cameroon, Equatorial Guinea, and Gabon. We used geospatial modeling to map patterns of genomic turnover (i.e. the change in allele frequencies with geographic distance) and predict areas of elevated genomic vulnerability. P. auritus serves as an ideal model for examining the effects of environmental heterogeneity because it occurs in a variety of forest types (Zimkus & Schick, 2010)and occupies a wide range of environmental conditions. Findings from these methods were then used to address the following hypotheses: 1) Temperature and precipitation are the most important drivers of morphological differentiation 2) IBE will influence genomic differentiation more than IBB or IBP 3) Areas of greatest environmentally-associated genomic turnover are associated with strong environmental gradients across the landscape 4) Patterns of environmentally-associated genomic variation reflect those observed for phenotypic variation 5) Genomic vulnerability will be highest in areas expected to undergo the greatest environmental change.
2. MATERIALS AND METHODS