3.5 Patterns of genomic turnover and genomic vulnerability across environmental space
We used a GF approach to determine associations between SNPs and environmental variables and map environmentally-associated genomic turnover across the total study area. A total of 1396 SNPs (45% of all SNPs) had R2 values > 0 (0.01-0.704, average 0.33). When testing model performance, the number of SNPs with R2 values > 0 for all of the randomized datasets fell below the number observed for the real data (Fig. S2) and the mean R2 value generated for the real dataset fell within the upper 95% quartile of values generated for the randomized datasets (Fig. S3), both indicating that the GF model shows a stronger association between environmental and genomic variation for our dataset relative to the set of randomized datasets. Precipitation of the coldest quarter, precipitation of the warmest quarter, and latitude were the most important environmental predictors of genomic turnover (Fig. 3b). Projected associations between allele frequencies and precipitation of the coldest quarter, precipitation of the warmest quarter, and latitude revealed areas of pronounced genomic turnover across the Cameroonian highlands (pink to orange), forest-savanna ecotone of south-central Cameroon (orange to green) and across the equator (green to blue) (Fig. 3c). There was also a moderate gradient from the coast to the interior of Gabon (dark blue to light blue).
Predictions of environmentally associated genomic turnover under future climate change projections showed the same general pattern of genotype-environment associations across the landscape relative to current predictions (Fig. S4). However, there were notable differences in patterns of genomic turnover between present and future climate change scenarios that were relatively consistent across both RCPs. Subtracting the current predictions from the future predictions under climate change projections (RCP 2.6 & RCP 8.5) for the year 2080 revealed two distinct hotspots of genomic vulnerability (Fig. 3d & Fig. S5): one centered around the lower Sanaga basin and a second in the far southeastern region of Cameroon.
4. DISCUSSION
We adopted a comprehensive statistical approach to disentangling the effects of environment, geographic distance, and landscape barriers on phenotypic and genomic variation in the African puddle frog P. auritus . Overall, we find that environmental variation plays an important role in shaping patterns of morphological and genomic differentiation. This is in addition to, but independent of, geographic distance. In particular, seasonal patterns of temperature and precipitation appear to be key in driving patterns of diversification in this tropical region, in keeping with a recent meta-analysis conducted of environmentally-mediated selection across the tropics (Siepielski et al., 2017). We also find that environmentally heterogeneous landscapes are important generators of patterns of high phenotypic and genomic variation suggesting that they may play an important role in promoting and maintaining biodiversity.
Our first hypothesis posited that temperature and precipitation are the most important drivers of morphological differentiation. We found that temperature evenness and precipitation seasonality were significant predictors of all measures of phenotypic variation. Both temperature and precipitation are known to be important factors influencing amphibian development, growth, and population dynamics and are expected to be key determinants of survival under climate change (Ficetola & Maiorano, 2016; Pounds et al., 1999). Our linear regression results show that body size increases with temperature evenness and decreases with precipitation seasonality. Thus, larger bodies are expected to be found in more uniform habitats with less variation in temperature and precipitation. This is somewhat consistent with the converse water availability hypothesis such that in areas with less intense wet and dry periods there is likely more water available year-round, allowing for investment in growth.
Body size, relative leg length, and head shape are predicted to vary along the transition zone between forest and savanna in central Cameroon where patterns of elevated morphological divergence have also been reported in a sunbird (Smith et al., 2011). The Cameroonian highlands and coastal regions of the Gulf of Guinea also appear to be associated with variation in body size and head shape, possibly due to strong ecological gradients associated with elevation and precipitation in these regions. Differences in skull shape morphology have been linked to the type, size, and speed of prey consumed in frogs and other amphibians (Emerson, 1985; Kaczmarski et al., 2017; Van Buskirk & Schmidt, 2000; Vega-Trejo et al., 2014) indicating that head shape might be at least partly adaptive. Head morphology has also been shown to exhibit considerable developmental plasticity in response to changes in temperature and could have important consequences for post-larval survival (Tejedo et al., 2010). There are also strong gradients predicted in both body size and relative leg length differentiation from the coast to the interior of Gabon, likely influenced by the degree of variation found in the population from Gamba. These frogs may be an example of cryptic speciation considering they have smaller body sizes but larger relative leg lengths compared to the rest of the samples and also given their high levels of genomic differentiation from other sites. There is also evidence that P. auritus exhibits complex patterns of spatial niche partitioning (Zimkus et al., 2010) and extraordinary patterns of diversification (Gvoždík et al., 2020). Although we cannot disentangle the effects of genetic adaptation and phenotypic plasticity, it is important to note that tropical ectotherms are considered to be particularly sensitive to changes in temperature and/or precipitation so that even subtle shifts in these variables could have profound impacts on fitness (Deutsch et al., 2008; Ficetola & Maiorano, 2016).
Our second hypothesis stated that IBE will influence genomic differentiation more than IBB or IBP. Contrary to many phylogeographic studies that have been carried out previously in central Africa, we did not find evidence for an effect of landscape barriers or Pleistocene refugia on population genomic differentiation. These findings are in stark contrast to many previous studies that have placed emphasis on the role of Pleistocene refugia and/or rivers (Anthony et al., 2007; Bohoussou et al., 2015; Eriksson et al., 2004; Nicolas et al., 2011) with the exception of Bell et al. (2017) where rivers were not important in reed frog diversification. However, our findings provide strong support for the role of environment, specifically seasonal variation in patterns of precipitation, as the most important environmental factor. Geographic distance is also consistently identified as a strong predictor of genomic differentiation. The role of IBD was supported by findings from our Mantel tests, the significance of geographic distance in GDM, and the significance of latitude, but not longitude, in predicting genomic turnover in GF analyses.
Patterns of environmentally-associated genomic differentiation reported here are consistent with previous investigations of gene-environmental associations in this region. For example, precipitation has been shown to be an important predictor of patterns of genetic variation in central African lizards (Freedman et al., 2010), chimpanzees (Mitchell et al., 2015), birds (Smith et al., 2011), and forest antelope (Ntie et al., 2017). In the present study, precipitation of the coldest quarter is highest in the Cameroon highlands, and decreases progressively throughout central Cameroon and Gabon (Fig. S6a), mirroring shifts in genomic turnover observed in P. auritus. Conversely, precipitation of the warmest quarter is highest in most of Gabon, especially along the coast and decreases towards Cameroon (Fig. S6b). Both of these patterns demonstrate shifts in genomic differentiation throughout the highlands, across the equator, and subtly from coastal to inland Gabon. Gradients in rainfall not only shape the distribution of forest cover but also present potentially strong selection pressures on the phenology of P. auritus since the timing and duration of amphibian reproductive events are very sensitive to rainfall levels (Corn, 2005; Ficetola & Maiorano, 2016).
GDM also identified precipitation seasonality as a significant predictor of genomic turnover. This environmental variable is linked to seasonal patterns in rainfall availability that are inverted across the Equator separating Cameroon and Gabon. Rainforests either side of the equator have their own distinct seasonal patterns of rainfall (Heuertz et al., 2014) such that the dry season in central Cameroon coincides with the rainy season in northern Gabon and vice versa. This seasonal inversion could be responsible for the shift in genomic variation observed inP. auritus across the equator. It has been hypothesized that these contrasting patterns of seasonal rainfall could lead to reproductive isolation and speciation across this region (Heuertz et al., 2014). Future work should look more closely at the seasonal inversion hypothesis and how heterogeneous annual patterns of rainfall influence genomic differentiation in other rainforest species.
Our third hypothesis was that areas of greatest environmentally-associated genomic turnover are associated with strong environmental gradients across the landscape. Areas of elevated genomic turnover in P. auritus appear to correspond to known ecological gradients. Genomic turnover is predicted to be high throughout the forest-savanna ecotone region south of the montane region in Cameroon where rainforest habitat in the south gradually transitions to savanna in the north. These findings are consistent with patterns of high intraspecific genomic diversity across this ecotonal region in the rainforest bird Andropadus virens (Zhen et al., 2017) and soft-furred mouse Praomys misonnei (Morgan et al., 2020). There is also high genomic turnover in P. auritus across the Cameroon highlands, reflecting both elevation and distance from the coast. The Cameroon highlands are a known biodiversity hotspot, especially for amphibian richness and endemism (Amiet, 2008; Herrmann et al., 2005; Pauwels & Rodel, 2007; Zimkus & Gvoždík, 2013) so that elevated genomic turnover in this region is to be expected. Mountain ranges and elevational gradients are often recognized as important drivers of genetic heterogeneity and, as is the case here, are important for the conservation of evolutionary potential. Overall, our results show a strong role for environment in shaping genomic differentiation such that areas of elevated genomic turnover span regions of strong ecological transition, providing further support for the role of environmental gradients and ecotones in shaping adaptive diversification.
Our fourth hypothesis stated that patterns of environmentally-associated genomic variation reflect those observed for phenotypic variation. Patterns of environmentally-associated morphological and genomic variation are relatively similar in Cameroon. RF and GF projections suggest that the forest-savanna ecotone in Cameroon is predicted to result in elevated morphological and genomic variation. These projections are based on the aggregated effects of the significant environmental variables which are primarily related to seasonality. Western Cameroon is characterized by more densely forested areas with pronounced precipitation seasonality, which then transitions south to habitats that include both forest and savanna and experience especially high seasonal variability in temperature and precipitation (Sesink Clee et al., 2015; Smith et al., 2011). Our findings are consistent with previous examples indicating that seasonality in moisture levels and precipitation are key explanatory variables for both morphological and genomic variation within this region (Smith et al., 2011), and thus provides further support for the role of environmental variation in driving diversification. In Gabon, we find relatively uniform patterns of genomic variation relative to patterns of morphological variation. While seasonal variation is less pronounced relative to Cameroon, Gabon harbors a variety of heterogeneous habitats, such as narrow, coastal alluvial plains, extensive wetlands, patches of savanna, and low elevation mountain zones (Lee et al., 2006), many of which may present unique selection pressures contributing to phenotypic variation.
Finally, we posited that genomic vulnerability is predicted to be highest in areas expected to undergo the greatest environmental change. We identified several areas of high genomic vulnerability where populations may be more susceptible to climate change under future projections. In the present study, the Sanaga River delta area in southwest Cameroon and an area in the southeast of the country, north of Lobéké and Nki National Parks, are predicted to be regions of greatest genomic vulnerability. While most of the Sanaga River is unprotected, the Douala Edea Wildlife Reserve falls within this area and constitutes an important target for continued protection. Genomic vulnerability may be an important metric to incorporate into conservation prioritization as it may also indicate areas where populations are already susceptible to present-day environmental pressures. For example, Bay et al. (2018) have recently shown that yellow warbler (Setophaga petechia ) populations with the highest genomic vulnerability were also experiencing the largest population declines. Therefore, areas of high genomic turnover and vulnerability may be important targets for future conservation efforts since the former serves as centers of high adaptive potential whereas the latter signal susceptibility to environmental change.
Although we adopted a genome-wide approach in the present study, our SNP dataset is only likely to capture a fraction of the total number of loci in the genome that constitute targets for selection and/or regions of the genome that may be linked loci under selection. Further research should focus on linking genotypic variation to phenotypic traits under selection to more fully understand the evolutionary significance of divergence across ecological gradients as well as examine the relative importance of genetic versus environmental factors in contributing to the observed morphological variation.
Understanding the ecological and historical processes involved in diversification is important not only for increasing our knowledge of evolutionary mechanisms, but also for making evolutionarily informed conservation decisions to protect biodiversity and prioritize new area for preservation in the light of rapid climate change. By taking a robust statistical approach to disentangling competing drivers of differentiation, we show that environmental factors rather than historical barriers to gene flow are largely responsible for patterns of morphological differentiation and genomic turnover in our study species. These findings, therefore, highlight the importance of preserving heterogeneous environments, such as environmental gradients, in maintaining species adaptive evolutionary potential and underline the importance of considering evolutionary processes in the design of future protected areas.
ACKNOWLEDGEMENTS
We thank the Agence Nationale des Parcs Nationaux, ANPN (permit #AE130012), Centre National de la Recherche Scientifique et Technologique, CENAREST (permit #AR0010/13, AR0024/14), Ministère des Forêts et de la Faune, MINFOF (permit #153/AO/MINFOF/PNCM, 008/A/MINFOF/R), and Ministère de la Recherche Scientifique et de l’Innovation, MINRESI, as well as all our valuable field guides for helping organize field collections and processing samples for exportation. We also thank University of California, Berkeley’s Vincent J. Coates Genomic Sequencing Laboratory (GSL), for sequencing services. This research was supported by National Science Foundation grant OISE 1243524. Finally, we would like to thank the two anonymous reviewers for their extensive suggestions and comments which led to the improvement of this manuscript.
REFERENCES
Abernethy, K. A., Coad, L., Taylor, G., Lee, M. E., & Maisels, F. (2013). Extent and ecological consequences of hunting in Central African rainforests in the twenty-first century. Philosophical Transactions of the Royal Society B: Biological Sciences ,368 (1625).
Aleixo, A. (2004). Historical diversification of a terra-firme forest bird superspecies: A phylogeographic perspective on the role of different hypotheses of amazonian diversification. Evolution ,58 (6), 1303–1317.
Ali, O. A., O’Rourke, S. M., Amish, S. J., Meek, M. H., Luikart, G., Jeffres, C., & Miller, M. R. (2016). RAD capture (Rapture): flexible and efficient sequence-based genotyping. Genetics , 202 (2), 389–400.
Amiet, J. (2008). Frog Diversity in Cameroon. In S. Stuart, M. Hoffmann, J. Chanson, N. Cox, R. Berridge, P. Ramani, & Y. BE (Eds.),Threatened amphibians of the world. Lynx Edicions with IUCN—The World Conservation Union, Conservation International and NatureServe(p. 16).
Anthony, N. M., Atteke, C., Bruford, M. W., Dallmeier, F., Freedman, A., Hardy, O., Ibrahim, B., Jeffery, K. J., Johnson, M., Lahm, S. A., Lepengue, N., Lowenstein, J. H., Maisels, F., Mboumba, J. F., Mickala, P., Morgan, K., Ntie, S., Smith, T. B., Sullivan, J. P., … Gonder, M. K. (2015). Evolution and Conservation of Central African Biodiversity: Priorities for Future Research and Education in the Congo Basin and Gulf of Guinea. Biotropica , 47 (1), 6–17.
Anthony, N. M., Johnson-Bawe, M., Jeffery, K., Clifford, S. L., Abernethy, K. A., Tutin, C. E., Lahm, S. A., White, L. J. T., Utley, J. F., Wickings, E. J., & Bruford, M. W. (2007). The role of Pleistocene refugia and rivers in shaping gorilla genetic diversity in central Africa. Proceedings of the National Academy of Sciences of the United States of America , 104 (51), 20432–20436.
Balkenhol, N., Dudaniec, R. Y., Krutovsky, K. V, Johnson, J. S., Cairns, D. M., Segelbacher, G., Selkoe, K. A., von der Heyden, S., Wang, I. J., Selmoni, O., & Joost, S. (2017). Landscape Genomics: Understanding Relationships Between Environmental Heterogeneity and Genomic Characteristics of Populations. In O. P. Rajora (Ed.), Population Genomics: Concepts, Approaches and Applications (pp. 261–322). Springer International Publishing.
Bay, R. A., Harrigan, R. J., Underwood, V. Le, Gibbs, L., Smith, T. B., & Ruegg, K. (2018). Genomic signals of selection predict climate-driven population declines in a migratory bird. Science ,361 (6401), 83–86.
Bell, R. C., Parra, J. L., Badjedjea, G., Barej, M. F., Blackburn, D. C., Burger, M., Channing, A., Dehling, J. M., Greenbaum, E., Gvoždík, V., Kielgast, J., Kusamba, C., Lötters, S., McLaughlin, P. J., Nagy, Z. T., Rödel, M. O., Portik, D. M., Stuart, B. L., VanDerWal, J., … Zamudio, K. R. (2017). Idiosyncratic responses to climate-driven forest fragmentation and marine incursions in reed frogs from Central Africa and the Gulf of Guinea Islands. Molecular Ecology , 26 (19), 5223–5244.
Boaratti, A. Z., & Da Silva, F. R. (2015). Relationships between environmental gradients and geographic variation in the intraspecific body size of three species of frogs (Anura). Austral Ecology ,40 (8), 869–876.
Bohoussou, K. H., Cornette, R., Akpatou, B., Colyn, M., Kerbis Peterhans, J., Kennis, J., Šumbera, R., Verheyen, E., N’Goran, E., & Katuala, P. (2015). The phylogeography of the rodent genus Malacomys suggests multiple Afrotropical Pleistocene lowland forest refugia.Journal of Biogeography , 42 (11), 2049–2061.
Born, C., Alvarez, N., Doyle, M., Ossari, S., Wickings, E. J., Hossaert-McKey, M., & Chevallier, M.-H. (2011). Insights into the biogeographical history of the Lower Guinea Forest Domain: evidence for the role of refugia in the intraspecific differentiation of Aucoumea klaineana. Molecular Ecology , 20 (1), 131–142.
Castellano, S., & Giacoma, C. (1998). Stabilizing and directional female choice for male calls in the European green toad. Animal Behaviour , 56 (2), 275–287.
Catchen, J., Hohenlohe, P. A., Bassham, S., Amores, A., & Cresko, W. A. (2013). Stacks: An analysis tool set for population genomics.Molecular Ecology , 22 (11), 3124–3140.
Catchen, J. M., Amores, A., Hohenlohe, P., Cresko, W., Postlethwait, J. H., & De Koning, D.-J. (2011). Stacks: Building and Genotyping Loci De Novo From Short-Read Sequences. Genes Genomes Genetics ,1 (3), 171–182.
Charles, K. L., Bell, R. C., Blackburn, D. C., Burger, M., Fujita, M. K., Gvoždík, V., Jongsma, G. F. M., Kouete, M. T., Leaché, A. D., & Portik, D. M. (2018). Sky, sea, and forest islands: Diversification in the African leaf-folding frog Afrixalus paradorsalis (Anura: Hyperoliidae) of the Lower Guineo-Congolian rain forest. Journal of Biogeography , 45 (8), 1781–1794.
Colyn, M., Gautier-Hion, A., & Verheyen, W. (1991). A Re-Appraisal of Palaeoenvironmental History in Central Africa: Evidence for a Major Fluvial Refuge in the Zaire Basin. Journal of Biogeography ,18 (4), 403–407.
Corn, P. S. (2005). Climate change and amphibians. Animal Biodiversity and Conservation , 28 (1), 59–67.
Cushman, S. A., & Schwartz, M. K. (2006). Gene Flow in Complex Landscapes: Testing Multiple Hypotheses with Causal Modeling.