Interview with the authors: testing the role of ecological selection on color pattern variation

The color variation that exists among individuals has lent itself to the study of selection since Darwin. Recently, Zaman, Hubert, and Schoville (2019) investigated the effects of selection on the diversity of the wing color pattern in the butterfly Parnassius clodius across a large portion of its range. These researchers found evidence supporting the idea that coloration may serve as a warning signal to predators, providing some predator avoidance benefits to individuals. In addition, the variation of solar radiation and precipitation observed geographically across sites was negatively correlated with the amount of melanin observed at each site. This suggests that the occurrence of melanin may provide a selective advantage in the form of thermoregulatory function. For more information on how selection influences butterfly wing coloration, please see the full article and the interview with Dr. Schoville below. 

Parnassius clodius mating, male located below female. Photo by Sean Schoville.

What led to your interest in this topic / what was the motivation for this study? Khuram and I were both interested in butterfly color pattern variation, and in particular, cases where there might be competing selective pressures acting on wing pattern phenotypes. Most work on butterfly wing patterns focuses on predator-prey interactions and aposematic colors, but butterfly wings are essential to flight performance and important in thermal regulation. A number of recent papers have shown that butterfly color pattern appears to be responding to climate warming, and then there are well-known cases (such as alpine Colias, thanks to Ward Watt) where thermoregulation has been linked to basking behavior and the pigment on wings. Thus, in examining variation in Parnassius clodius (which occurs over a broad elevation and latitudinal range), we hoped that we could decouple environmental signals that might act on different wing color elements. 

What difficulties did you run into along the way? Sampling our butterflies across this large region was a major challenge, particularly as adult flight times are rather short (two to three weeks). And then we were surprised by the strong difference in adult phenology across sites (some adults are active in May, others in late July). This is evident regionally (Utah versus Washington), and across elevation within a region. In the end, it required three years of effort, with some very long road trips from Wisconsin.

What is the biggest or most surprising innovation highlighted in this study? After some initial efforts to link aposematic variation (red eyespots) indirectly to predator communities (through climate variables that might covary with predator abundance), we realized this was too tenuous. So, we were delighted to discover publicly available data on bird abundance. While this did not solve the problem (perhaps due to lack of spatial resolution in the bird data), I think using this data to analyze butterfly wing patterns was one of the more innovative aspects of our paper. We had a much easier time linking spatial climate data to melanization (dark pigmentation). As an aside, this raises the important point that some data, i.e. abiotic environmental data, is much easier to come by than biotic data. This is unfortunate, as we expect biotic selective forces to be equally or more important drivers of microevolution in some cases.

Lead author on this study, Khuram Zaman, after a day of collecting samples. Photo by Sean Schoville.

Moving forward, what are the next steps in this area of research? We’d like to extend this work in two directions. First, we’d like to connect color pattern traits to underlying genes and measures of heritability. Although the genes controlling butterfly color pattern are well studied, to date no representative of the snow Apollo subfamily Parnassinae have been included in these efforts. Members of the family are tremendously variable and quite stunning in their dramatic contrasts of color. Second, experiments are needed to link our inferences of ecological selection to fitness differences, as well as performance in the field. Physiological assays of melanic variants, coupled with mechanistic thermodynamic models, have been developed for Colias butterflies (see Joel Kingsolver and Lauren Buckley’s work). This type of modeling could provide important connections to conservation of Parnassius clodius populations under changing climates, and might perhaps extend to conservation work on other highly threatened Parnassius species.

What would your message be for students about to start developing or using novel techniques in Molecular Ecology? The development of novel approaches is a key part of advancing biological knowledge, but it can be a daunting endeavor given the breadth and scope of the scientific literature nowadays. Integrating multiple approaches, on the other hand, can equally help to advance our knowledge and provide opportunities to address long-standing questions. This is the direction we took in this paper. My personal view is that students should to try to master multiple techniques (assemble a toolkit, so to speak) and apply those techniques to fundamental problems. Hopefully, it’s a lot of fun in the process and leads to interesting collaborations!

What have you learned about methods and resources development over the course of this project? We are entering a golden age of data-rich resources, in terms of spatial environmental data and genomics data. These increasingly provide the power to test refined hypotheses about evolutionary and ecological processes, and are becoming more accessible to all researchers. One of my favorite accomplishments in the paper is using genetic covariance data among populations (relatedness data) as a covariate in fitting morphology ~ environment models. The use of such population contrasts is important in controlling for non-independence in the data due to ancestry. While we have known about the importance of genetic covariance in hypothesis testing for some time (thanks to Joseph Felsentstein’s work), it is only recently possible to use genome-wide data. This provides very precise measures that are highly informative, and enabled us to rule out the role of genetic drift as a driver of wing pattern variation. 

Describe the significance of this research for the general scientific community in one sentence. Our research demonstrates that butterfly wing color patterns evolve in response local climate conditions, as a way to regulate body temperature.

Describe the significance of this research for your scientific community in one sentence. Our work demonstrates that elements of butterfly wing pattern phenotypes respond independently to different sources of selection, with climate variation acting on thermoregulatory ability as an important driver of butterfly color pattern.

Parnassius clodius basking to thermoregulate. Photo by Sean Schoville.

Zaman K, Hubert MK, Schoville SD. 2019. Testing the role of ecological selection on color pattern variation in the butterfly Parnassius clodius. Molecular Ecology 28:50586-5102.

Interview with the authors: Environmental heterogeneity and not vicariant biogeographic barriers generate community‐wide population structure in desert‐adapted snakes

Phylogeographic studies have long focused on striking biogeographic barriers, and comparative phylogeography often looks for shared divergence across such barriers as evidence of shared responses to similar environments across taxa. However, in addition to such barriers, geographic distances and local adaptation to environmental heterogeneity may shape genetic divergence. In their recent Molecular Ecology paper, Myers and colleagues collect genomic data from 13 co-distributed species of snakes from Southwestern North America and evaluate the relative importance of biogeographic barriers, geographic distance, and environmental heterogeneity in structuring genetic divergence. Much of the previous phylogeographic work in this region has focused on divergence across a prominent biogeographic barrier: the Cochise Filter Barrier (CFB), which separates the Sonoran and Chihuahuan Deserts, and divergence across this barrier has been suggested to be an important factor driving divergence in snakes from the region. Though they expected to find a prominent role of this barrier, instead, Myers and colleagues find strong support for geographic distance and environmental heterogeneity as important factors structuring genetic divergence, but less support for biogeographic barriers. Further, they find that different variables contribute most to divergence across the 13 taxa studied, highlighting the importance of species-specific responses to environmental variation. Read the full article here, and read below for a behind-the-scenes interview with lead author Edward Myers.

What led to your interest in this topic / what was the motivation for this study? 
As a research team we have a general interest in what factors are promoting population genetic differentiation and whether codistributed species have similar evolutionary histories in response to shared environmental changes over time. Specifically, in this system where there is a well known biogeographic barrier (Cochise Filter Barrier; CFB), we were interested in whether entire assemblages of taxa show similar population structure. Initially the motivation for this study was to assess the degree of co-divergence across the CFB, however, as we analyzed these data it became clear that we needed to incorporate spatial and environmental data to understand population divergence. This study also allowed me spend a significant amount of time in the field collecting tissue samples from snakes!

What difficulties did you run into along the way? 
One of the biggest difficulties with this study was handling and analyzing all the generated data. We had almost 400 samples sequenced for radseq, so processing and analyzing these data took a significant amount of computational time. Also, one difficulty was the logistics of collecting fresh tissue samples for all of these species across the southwestern US and northern Mexico, but issues like this are easily over come by collaborating.

What is the biggest or most surprising finding from this study? 
The biggest surprise from this study is that patterns of isolation-by-distance and isolation-by-environment are more important in explaining population genetic differentiation than a commonly cited biogeographic barrier. This result really stresses the importance of incorporating spatial analyses when analyzing phylogeographic data because aspatial analyses may result in spurious results of population structure and mislead our ideas of what is driving population divergence and speciation.

Moving forward, what are the next steps for this research? 
Moving forward I plan to generate whole genome sequence data for species within this system to understand what loci may be under selection in response to environmental heterogeneity. Given the strong signature of IBE I expect to find patterns of strong selection along transects of temperature and precipitation across the Sonoran and Chihuahuan Deserts. Further, I am interested in how other regions globally that have been cited as important biogeographic barriers in phylogeographic studies might also be strongly influenced by patterns of IBD and IBE, and not vicariant barriers.

What would your message be for students about to start their first research projects in this topic? 
There is so much great work published in the field of landscape genetics and comparative phylogeography and I would suggest that students start by combing through that work first. But as general advice I would suggest that students really explore their data in a meaningful way and spend some time thinking about what factors could be responsible for similar patterns observed in a genomic data set (e.g., IBD vs vicariance or selection vs historical demography).

What have you learned about science over the course of this project? 
I have really learned that genomic data should be carefully analyzed as to not be influenced by preconceived ideas of the system that you might be working within. Also, I think that this is becoming more and more true, but you have to collaborate in order to do great science.

Describe the significance of this research for the general scientific community in one sentence.
This work demonstrates that codistributed species do not have shared evolutionary histories, and that they do not respond to the same landscape and shared environment in similar ways.

Describe the significance of this research for your scientific community in one sentence.
Our work shows that simple patterns of isolation-by-distance and isolation-by-environment have contributed to population genetic differentiation more so than commonly cited biogeographic barriers.

Full article: Myers EA, Xue AT, Gehara M, et al.Environmental heterogeneity and not vicariant biogeographic barriers generate community‐wide population structure in desert‐adapted snakes. Mol Ecol. 2019;28:4535–4548. https://doi.org/10.1111/mec.15182

Interview with the authors: quality and quantity of genetic relatedness data affect the analysis of social structure

Understanding the influence of relatedness on fine-scale social interactions within a population is fundamental to understanding the role of kinship in animal societies. In this study, Foroughirad et al provide insight into the quality of Single Nucleotide Polymorphism (SNP) data required to obtain accurate and precise parentage assignments and relatedness coefficients using data from a long‐term behavioural study on bottlenose dolphins with a known partial pedigree. They then go on to explore how the quality of these estimates influence post-hoc analyses exploring the relationship between relatedness and social structure. Again, they provide important practical guidance about the quality of data needed for these types of analyses. This article was published in Molecular Ecology Resources: read the full article here, and read our interview with Vivienne Foroughirad, lead author of the study, below.

An adult female bottlenose dolphin with her six-month old calf. Photo Credit: Vivienne Foroughirad

What led to your interest in this topic / what was the motivation for this study? 

In the broadest sense my research interests concern the evolution of sociality and complex social behaviors such as cooperation. To that end, I was interested in our ability to parcel out contexts in which cooperation occurs between kin versus between unrelated individuals. Non-kin cooperation is rare is animal societies, and a common way to search for examples is to first investigate the link between the strength of social relationships and the genetic relatedness of pairs. Since genotyping-by-sequencing is now cheaper and more accessible than ever, I wanted to explore the effects this increased resolution would have on our power to test the relationship between social structure and relatedness, especially in viscous populations with strong philopatry.

What difficulties did you run into along the way? 

In our case, the greatest challenges centered around maintaining a longitudinal study on a wild marine mammal with a large enough sample size to make answering these types of questions feasible. We were lucky to have over 30 years of data available from the Shark Bay Dolphin Project which allowed us to verify some of the reconstructed pedigree relationships, as well as measure detailed home range usage and social associations. An analytical difficulty we encountered is how to account for the confounding effect of philopatry or limited dispersal on social relationships with kin if you want to distinguish kin discrimination from more passive kin associations that are a byproduct of shared space use. 

What is the biggest or most surprising finding from this study? 

We provided evidence that genotyping-by-sequencing methods could produce more precise relatedness values than typical microsatellite analyses, which isn’t surprising. What was less well-understood was the effect this would have for downstream analyses, such as those testing whether relatedness correlated with social affiliation. We found that even though our study species exhibits strong, life-long affiliative relationships between maternal kin, there were a surprising number of scenarios under which our analysis failed to detect a significant correlation between genetic relatedness and social associations. We also found surprisingly diminishing returns in relatedness resolution with increasing sample size (number of individuals) when small numbers of markers were used. 

Moving forward, what are the next steps for this research? 

Pedigree reconstruction is rapidly improving, especially where there is access to new genetic resources such as chromosome-level assemblies for non-model organisms. Improved kin assignment methods will allow us to investigate the function of these relationships at the level of the individual, which will help us to tease out how both intra- and inter-specific variation in ecology and demography affect social behavior. Within my own study site, I’m using these data to look at the effect of family size on social network position and reproductive success, as well as the demographic conditions that facilitate the formation of non-kin bonds. We’re also working on ways to better discriminate between maternal and paternal kin, which will be important for investigating the mechanisms of kin recognition.

What would your message be for students about to start their first research projects in this topic? 

That this is a great idea! Rapid advances in technology will open up new avenues of inquiry and there is lots of work to be done. Nevertheless, as with any field, you also need to know when to stop and submit. There will always be a new higher coverage genome or updated version of the software you’re using that’s about to be released, but if you keep reanalyzing your data with each advance, you’ll never finish a project. My second piece of advice would be to practice simulating data and analyzing it. Building simulated datasets, tweaking parameters, and testing different software has really deepened my understanding of methodologies- plus you can start before you even get your first sequencing results and be ready with a tested pipeline when you do get results in hand.

What have you learned about science over the course of this project? 

That building a robust, reproducible, and well-documented pipeline for analysis is crucial. It might take a bit more work to set up, but it’s always worth it. I also benefitted a lot from the opportunity to present my work to audiences from different disciplines which helped me keep the big picture in mind since I’m the kind of person that gets easily caught up in minutiae. Biologically, I’m always reminded that there’s so much individual variation that gets masked by conducting analyses at the population level, and that rather than being discounted as noise, that variation could be leveraged to ask really interesting questions about how ecology and demography affect behavior. 

Describe the significance of this research for the general scientific community in one sentence.

The correlation between genetic relatedness and the strength of social relationships can be masked by the limited power of typical published sample sizes.

Describe the significance of this research for your scientific community in one sentence.

We provide practical guidance for how sample sizes and sequencing methods might interact to improve precision of relatedness estimates and their effect on the analysis of social structure, using wild bottlenose dolphins as a case study.

Three juvenile male bottlenose dolphins surface synchronously. Photo Credit: Vivienne Foroughirad

Foroughirad, V., Levengood, A. L., Mann, J., & Frère, C. H. (2019). Quality and quantity of genetic relatedness data affect the analysis of social structure. Molecular Ecology Resources, 1181–1194. https://doi.org/10.1111/1755-0998.

Interview with the authors: Linking plant genes to insect communities: Identifying the genetic bases of plant traits and community composition

Much research in community genetics attempts to understand how genetic variation influences community composition, but the majority of studies have been done at the level of the genotype. In their new Molecular Ecology paper, Barker and colleagues use genome-wide association mapping in aspen (Populus tremuloides) to identify specific genes that may influence variation in tree traits or in insect communities. They uncover 49 SNPs that are significantly associated with tree traits or insect community composition. Notably, insects with closer associations with host plants have more genetic correlations than less closely associated insects. Barker and colleagues find a SNP associated with insect community diversity and the abundance of interacting species, providing a link between genetic variation in aspen and insect community composition. Finally, they find that tree traits explain some of the significant relationships between SNPs and insect community composition, suggesting a mechanism by which these genes may influence community composition. Read the full article here, and get a behind-the-scenes interview with lead author Hilary Barker below.

What led to your interest in this topic / what was the motivation for this study? 
For some time, we have been interested in extended phenotypes – the idea that the genes of an organism not only shape the immediate traits of that organism, but also extensions of these traits, such as the community of insects living on a tree. Yet, until our study, most of the previous research had been largely focused on differences across genotypes of ‘host’ organisms (e.g., aspen, cottonwoods, evening primrose), rather the underlying genes. Thus, there were a lot of unknowns yet to be discovered. For instance, would the genetic effects be large enough to detect and identify? Would more underlying tree genes be found for insects that are more closely associated with the tree (i.e., leaf gallers) rather than free feeding insects? Would there be an overlap between genes associated with insect communities and genes associated with particular tree traits? 

What difficulties did you run into along the way? 
I think the largest challenge of conducting a Genome-Wide Association study on a common garden of trees is the planting and maintenance of this small forest. We had 1824 trees that needed planting, phenotyping, and care. This work was most intense in the first four years of the study to ensure that each tree survived a summer drought and multiple harsh winters. The next most challenging hurdle was conducting the insect surveys. These surveys involved a large team effort and happened during some of the hottest days of the summer. 

What is the biggest or most surprising finding from this study? 
The most exciting finding from this study was the identification of an aspen gene (early nodulin-like [ENODL] transmembrane protein, Potra001060g09097) that underlies insect community composition; both diversity and the abundance of key insect species (aphids and ants). While we do not yet know the mechanism by which this gene influences insect communities, we do know that this protein is involved in the transportation of carbohydrates. Thus, it’s possible that this gene directly influences aphids and ants via their interactions with carbohydrate-rich honeydew, and/or indirectly influence insects via numerous tree traits, including both growth (size) and defense. To our knowledge, this is the first identification of allelic variation in a plant gene that is associated with a complex insect community trait (i.e., insect community composition).

Moving forward, what are the next steps for this research? 
The next step of this research is to explore how the genetic underpinnings of these aspen traits and associated insect communities may vary across different environmental gradients and with tree ontogeny. Previous research has shown that aspen growth and defense traits vary with tree age, and these traits play a significant role in determining which insects will feed upon the foliage. Thus, the genetic contributions of insect community composition may vary substantially for more mature trees. The Lindroth lab is currently working on an expanded version of this study with more detailed traits and mature (reproductive) trees. In addition, gene expression will vary with different environmental conditions, which will likely also modify which genes are most important in shaping insect communities.

What would your message be for students about to start their first research projects in this topic? 
To complete a large community genomics study such as this, you will need a few key things. First, you will need a lot of help. Start recruiting anyone and everyone in sight. Mentoring undergraduates will be essential and ensuring that you can effectively asses the learning of your mentees and volunteers is critical (e.g., can they correctly identify X insect? Can they successfully complete X protocol in the lab?). Second, get organized. Project management platforms can be really helpful (e.g., Asana, MS Teams, etc.) to keep track of tasks. Third, refine your R markdown scripts. You will generate more data than you know what to do with, and thus creating R scripts to clean up, organize, and analyze your data will be a top priority. Also, if you can get a digital microscope (e.g., Dino Lite), then the tedious task of keying out insect specimens will be much easier and less cumbersome! I highly recommend it.

What have you learned about science over the course of this project? 
In terms of a genome-wide association study, it is best to have as large a sample size as possible (more genotypes and genetic variation). You do not want to invest a lot of resources into a study that has low statistical power for association testing. Also, phenotype as many traits as you can. At the onset, it is impossible to know which genes, if any, will be associated with which traits. Thus, you could end up with a lot of investment while identifying a small number of associated genes, or potentially no genes at all. 

Describe the significance of this research for the general scientific community in one sentence.
Our findings show that specific genes in a host organism can shape the composition of associated communities.

Describe the significance of this research for your scientific community in one sentence.
Complex extended phenotypes such as community composition have an identifiable genetic basis, and thus we can use this information to test and study the extent and limitations of community evolution.

Full article: Barker HL, Riehl JF, Bernhardsson C, et al. Linking plant genes to insect communities: Identifying the genetic bases of plant traits and community composition. Mol Ecol. 2019;28:4404–4421. https://doi.org/10.1111/mec.15158

Summary from the authors: City life alters the gut microbiome and stable isotope profiling of the eastern water dragon (Intellagama lesueurii)

Whilst urbanisation poses a major threat to many species, there is growing evidence to suggest that some species, labelled ‘urban adapters’, are thriving within the urban landscape. Urban landscapes differ drastically from native habitats, where urban adapters are often exposed to a more diverse range of novel food items compared to their rural counterparts, which frequently includes human subsidised resources. Diet is one of the most important factors influencing the gut microbiome, an extremely influential symbiotic community that plays a critical role in many processes affecting host health and fitness, including metabolism, nutrition, immunology and development. Here, using populations of the eastern water dragon in Queensland, Australia, we explore the link between urbanisation, diet and gut microbial changes. We show that city dragons exhibit a more diverse gut microbiome than their rural counterparts, and display microbial signatures of a diet that is richer in plant-material and higher in fat. Elevated levels of the Nitrogen-15 isotope in the blood of city dragons also suggests their diet may be richer in protein. These results highlight that urbanisation can have pronounced effects on the gut microbial communities of wild animals, but we do not yet know the possible repercussions of these microbial changes.

Full article: Littleford‐Colquhoun BL, Weyrich LS, Kent N, Frere CH. City life alters the gut microbiome and stable isotope profiling of the eastern water dragon (Intellagama lesueurii). Mol Ecol. 2019;28:4592–4607. https://doi.org/10.1111/mec.15240

Summary from the authors: Host plant associations and geography interact to shape diversification in a specialist insect herbivore

The sexual generation female of the cynipid gall wasp Belonocnema treatae, which forms locally adapted populations on three closely related live oaks across the southeastern United States. Photo credit: Jena Johnson.

In this study, we wanted to know how geography and ecology predicted population genetic structure among 58 populations of the gall wasp Belonocnema treatae, which exhibits regional specialization on three host plant species across the U.S. Gulf Coast. We combined range-wide sampling with a genotype-by-sequencing approach for 40,699 SNPs across 1,217 individuals. Disentangling the processes underlying geographic and environmental patterns of biodiversity is challenging, as such patterns emerge from eco‐evolutionary processes confounded by spatial autocorrelation among sample units. We evaluated this question using a hierarchical Bayesian model (ENTROPY) to assign individuals to genetic clusters and estimate admixture proportions. Using distance-based Moran’s eigenvector mapping, we generated regression variables that represent varying degrees of spatial autocorrelation in genetic variation among sample sites. These spatial variables, along with host association, were incorporated in distance-based redundancy analysis (dbRDA) to partition the relative contributions of host plant and spatial autocorrelation. This novel approach of combining ENTROPY results with dbRDA to analyze SNP data unveiled a complex mosaic of diversification within and among insect populations forming discrete host associated lineages coupled with geographic variation. This demonstrates that geography and ecology play significant roles in explaining patterns of genomic variation in B. treatae – an emerging model of ecological speciation.

Full article: Driscoe AL, Nice CC, Busbee RW,Hood GR, Egan SP, Ott JR. Host plant associations and geography interact to shape diversification in a specialist insect herbivore. Mol Ecol. 2019;28:4197–4211. https://doi.org/10.1111/mec.15220

Summary from the authors: Polygenic selection drives the evolution of convergent transcriptomic landscapes across continents within a Nearctic sister-species complex

Lake Whitefish (Coregonus clupeaformis)
Adult size benthic (normal) and limnetic (dwarf) sympatric whitefish
Sampled from Cliff Lake, Maine, USA.
Illustration by Paul Vecsei.

The Lake Whitefish (Coregonus clupeaformis) is found in Nearctic post-glacial freshwater lakes, and diverged about 500,000 years ago from its sister species, the European Whitefish (Coregonus lavaretus) found throughout Northern Europe, European Alpine lakes and Russia. These two lineages underwent an adaptive radiation following the last glaciation, resulting in the sympatric occurrence of limnetic and benthic species-pairs. Decades of research has aimed to decipher the process of adaptive divergence and ecological speciation in the Whitefish species complex. Here, we compared independent diverging species-pairs from the two continents to elucidate the genomic and transcriptomic bases associated with the benthic-limnetic diversification. We used a statistical framework to detect polygenic targets of selection associated with phenotypic diversification. We identified a subset of genes that showed convergent patterns of differential expression between limnetic and benthic species across both continents. Those adaptive divergent genes retained a higher degree of shared polymorphism among species-pairs, most likely due to balancing selection, and this genetic variation was associated with changes in levels of gene expression between species. As such, our results indicate that standing genetic variation underlying phenotypes involved in the ecological speciation of the whitefish species-pairs has been partly maintained in parallel across both continents for at least half a million years.

-Clément Rougeux – Postdoctoral Researcher, University of Calgary

Full article: Rougeux C, Gagnaire P‐A, Praebel K, Seehausen O, Bernatchez L. Polygenic selection drives the evolution of convergent transcriptomic landscapes across continents within a Nearctic sister species complex. Mol Ecol. 2019;28:4388–4403. https://doi.org/10.1111/mec.15226