Interview with the authors: Evaluation of model fit of inferred admixture proportions

Admixture models are widely-used in population genetics, but they make several simplifying assumptions, which, if violated, could result in misleading estimates of individual ancestry proportions. In a recent paper published in Molecular Ecology Resources, Garcia-Erill and Albrechtsen introduce evalAdmix, a program for detecting poor fit of admixture models to empirical data. evalAdmix uses the correlations of the residual differences between true and predicted genotypes to detect poor fit; when the assumptions of the model are not violated, the residuals of a pair of individuals should be uncorrelated. In simulation studies and analyses of empirical datasets, evalAdmix was useful in identifying model violations due to gene flow from unsampled ghost populations, continuous variation, population bottlenecks, and an incorrect assumed number of ancestral populations. Read the full article here, and read below for an exclusive interview with lead author Genís Garcia-Erill.

Full text: Garcia-Erill G. and Albrechtsen A. Evaluation of model fit of inferred admixture proportions. Mol Ecol Resour. 2020;20:936–949. https://doi.org/10.1111/1755-0998.13171.

Admixture model and evaluation with our method applied to worldwide human genetic variation. A. Admixture proportions inferred with ADMIXTURE assuming K=5 for all human populations from the 1000Genomes project. B. Evaluation of admixture model with the correlation of residuals performed with evalAdmix. Positive correlations are indicative of a bad model fit. The correlation of residuals shows that modelling with an ancestral population for each of the 5 major continental groups leads to a bad fit within most populations, and furthermore it gives additional information. For example we can see that the populations more genetically distant from the rest with which they are grouped, like Luhya in Webuye, Kenya (LWK) or Finish in Finland (FIN), have higher correlations of residuals, or it indicates the presence of substructure in some populations like the Gujarati Indians in Houston, TX (GIH).

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

The admixture model is one of the most used methods in population genetics, but it has already been known for some time that there are many potential issues with it. Specifically a recent study described very nicely different scenarios that can lead to wrong conclusions when applying the admixture model (Lawson et al. 2018). For example, they showed how multiple scenarios can lead to the same admixture results, and they also presented a method, badMixture, that can distinguish between those scenarios and evaluate model fit. However badMixture is quite difficult to apply, so we thought it would be interesting to develop an alternative method that could help in guiding the interpretation of admixture model results.

What difficulties did you run into along the way?

My background is in Biology and I had limited experience in computer science and statistics when I started with this project, so most of the difficulties were related to my learning how to work in these two disciplines. The method itself was relatively straightforward, but in order for it to work properly we needed to find a way to correct the bias caused by the frequency estimation. The frequency correction is only a small part of the main article, but it was where we put most of the work during the development of the method; that ended up as a few pages full of equations in the supplementary material. Another aspect where I had to put considerable effort was in making the implementation, since again I did not have much experience in developing software that would (hopefully) be used by other people. That made me consider things I would not usually think about.

What is the biggest or most surprising innovation highlighted in this study?

I think the method itself is the main result of the study. As I said there is already a method to evaluate the admixture model fit, badMixture. However that method is rarely used, because it requires performing additional analyses with CHROMOPAINTER and also requires having data with good enough quality to at least call genotypes. The method we present is more generally accessible since it is based on information unique to the admixture model itself, meaning one can directly apply it to any data set to which the admixture model has been applied. So it provides what we think is a simple way, both in the application and in the interpretation, to evaluate the admixture model results.

Moving forward, what are the next steps in this area of research?

There are several directions in which this work could be expanded. Something we already spent some time on is trying to develop a more firm theoretical foundation for the correlation of residuals as a measure of model fit, for example expressing it in terms of individual-specific Fst and the distance between the populations from which they are sampled, in a framework similar to that in Ochoa and Storey (2018). In the end we could not figure out the math and left it as a short mention in the discussion, but that would be something very nice to do. We also could not find a good way to use the residuals to develop some sort of measure of model fit at a purely individual level (instead of depending on the relationship between pairs of individuals, as it does right now), and that would also be very nice to do. Moreover, individual frequencies can also be calculated using principal component analyses, so this method could be expanded to work as an evaluation of a PCA as a description of population structure. Finally what we are looking forward to the most is to see how the method is applied to different datasets and how that helps gain new scientific insights.

What would your message be for students about to start developing or using novel techniques in Molecular Ecology? 

I am myself a student who has very recently started developing and using novel techniques in Molecular Ecology, so I am not sure if I have enough experience and perspective to give any useful advice. But based on my limited experience, I would say that it is important not to be afraid to jump into new areas or fields where we feel like we might have too limited experience, and that often what at first seems very difficult will become more and more accessible and doable as we work on it.

What have you learned about methods and resources development over the course of this project?

I started working on this study during my Master studies, so it has been one of my first research experiences. Basically all I know about method development I learned during the course of this project, from the more practical skills related to developing and implementing a method to how to explain it, and make it accessible to the community that might be interested in using it. I realized that this can actually be very important, since it will affect how many people end up using it. Also, as a user of bioinformatics methods I really appreciate when I use a new method if it is easy to use and does not create too many problems.

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

It is important to consider the assumptions of the methods we use, since relevant violations of the assumptions might result in misleading or even meaningless results.    

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

It makes it possible and easy to evaluate the model fit of the admixture model at the individual-level in almost any context in which the admixture model is currently used, so it can be applied before concluding a population is a mixture of others, or it can help to choose a meaningful number of ancestral populations.

References

Lawson, D. J., Van Dorp L., and Falush, D.. A tutorial on how not to over-interpret STRUCTURE and ADMIXTURE bar plots. Nature Communications 2018;9.1: 1-11.

Ochoa, A. and Storey, J. D. FST and kinship for arbitrary population structures I: Generalized definitions. BioRxiv 2016: 083915.

Interview with the authors: Applying genomic data in wildlife monitoring

Massive parallel sequencing has led to an explosion of sequence data in recent years. However, the methods used to obtain such data are usually high-cost and time-intensive, and often require high-quality samples. This creates limits as to whether and how well such data can be used by researchers working in applied conservation science. Here, we speak to Alina von Thaden about her recent study in Molecular Ecology Resources. Using European wildcats as a case study, Alina and co-authors present a relatively low-cost and time-efficient workflow for the development and optimisation of microfluidic SNP panels, which can be used to obtain SNP data from minimally invasive samples. Beyond outlining the workflow and its applications, they go so far as to estimate the costs of their pipeline, providing valuable practical information for conservation scientists. Read on for an in-depth view of this study.

Monitoring elusive European wildcats (Felis silvestris) is heavily reliant on noninvasively collected DNA samples. Photo credit: Annsophie Schmidt.

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

We are mainly working on genetic monitoring of large carnivores and most of our research is based on noninvasively collected wildlife samples such as hairs, faeces and saliva traces. The field demands for very fast and reliable genetic analyses of samples with degraded DNA. And since funding is generally sparse in applied conservation, our methods need to be cost-effective and suitable for high-throughput approaches.

Genomic tools, on the other hand, usually involve large amounts of data, complex bioinformatic pipelines and typically rely on samples with high-quality DNA. We have been looking into ways to combine the advantages of genomics with the challenges of conservation monitoring. For some years now, we have been working with microfluidic arrays combined with reduced SNP panels and wanted to share our experiences with other labs interested in applying them.

What difficulties did you run into along the way? 

Setting-up and optimizing methodological resources comes along with several challenges – but there is a lot to learn! Most important to me was to remain skeptical about the results and to constantly validate them through analyzing the data from several perspectives and with different software. The validation of the technology also took a lot of extra lab hours, but we are confident that the workflow and guidelines that we present now will save others a lot of hands-on time and costs when optimizing SNP panels for degraded samples.

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

First of all, after years of developing the framework, we applied it to a new SNP panel designed for dog-wolf hybridization assessment (to be published) and found that the lab work for generating a new ready-to-use marker panel took us only a few weeks. To see the approach being proved effective was great and encouraged us to share it with the community.

Secondly, a large proportion of noninvasively collected samples could be run without or with only very few genotyping errors as compared to more traditional microsatellite-based genotyping (see also von Thaden et al. 2017). This has direct implications for genotyping costs and thus promotes the broader establishment of a genomic technology in applied conservation.

Alina von Thaden collecting reference samples of European wildcat (Felis silvestris) for testing a newly developed SNP panel. Photo credit: Annsophie Schmidt.

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

One of our next steps is to apply the technology to historical samples from museum collections. Additionally, we are going to implement the SNP panel from our current paper in routine genetic monitoring of European wildcats in Germany.

We currently develop other reduced SNP panels for a variety of endangered species in our lab, such as dormice and European bison. Besides neutral variation, we also aim to integrate functional markers, such as SNPs associated with disease susceptibility.

Further, we will test alternative platforms that will allow generating larger SNP sets for degraded samples. Ultimately, our long-term goal is the effective implementation of an “applied genomic wildlife monitoring” approach.

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

Get in contact with other groups working in this area! Sharing ideas and experience really helps to shape your project and refine the aims of your research. Most people are very cooperative and happy to contribute or answer questions.

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

Perseverance and tenacity. When exploring new directions the research journey may well become bumpy and lead you somewhere else than you initially expected. But it’s worth it – keep your goal in mind and be ready to rethink your strategy.

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

Bridging the gap between genomics and applied conservation is a key prerequisite for effective wildlife management, especially in the light of rapid biodiversity declines.

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

We demonstrate how reduced SNP panels can be efficiently developed and optimized for genotyping based on degraded wildlife samples.

References

von Thaden, A., Cocchiararo, B., Jarausch, A., Jüngling, H., Karamanlidis, A. A., Tiesmeyer, A. … Muñoz-Fuentes, V. (2017). Assessing SNP genotyping of noninvasively collected wildlife samples using microfluidic arrays. Scientific Reports, 7, 83. https://doi.org/10.1038/s41598-017-10647-w

Full paper

von Thaden, A., Nowak, C., Tiesmeyer, A., Reiners, T. E., Alves, P. C., Lyons, L. A., … & Hegyeli, Z. (2020). Applying genomic data in wildlife monitoring: Development guidelines for genotyping degraded samples with reduced single nucleotide polymorphism (SNP) panels. Molecular Ecology Resources. https://doi.org/10.1111/1755-0998.13136

.

Interview with the authors: Which software is best to use for de novo assembly?

Reduced representation sequencing (e.g. RAD and GBS) is becoming ever more popular, but for species which lack a reference genome, little work has been done to assess which software may be best suited to building de novo assemblies from this data. Here, we speak to Melanie LaCava of the University of Wyoming about her recent Molecular Ecology Resources article, which explores the accuracy of de novo assemblies built by various software programs using DNA generated from double-digest libraries. Melanie and her co-authors found highly variable degrees of accuracy of assemblies built by six different software programs, and discuss which programs are best suited to this application. They also highlight the importance of optimising parameter settings within any given software. Read on to get a behind-the-scenes view of this study.

The completeness of assemblies in simulations of unmutated genomes (a, d), in simulations of an equal number of SNPs and indels (b, e), and simulations of 1–5 base pair indels (c, f). Values are reported for five assemblers: CDHIT (green), STACkS (blue), STACkS2 (purple), VelVeT (pink) and VSeARCH (orange). The hue of each color corresponds to the percent match parameter setting used in the assembly. For more information on this figure go to Figure 1’s caption here.

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

This study began as a research project in a graduate-level course on computational biology at the University of Wyoming led by the senior author on the paper, Alex Buerkle. Dr. Buerkle initiated the project and worked with the rest of the coauthors to pursue this de novo assembly software comparison. As reduced representation genotyping-by-sequencing has become more popular, new and repurposed software programs have been applied to each step in the bioinformatics pipeline. When a reference genome is unavailable for a study species, de novo assembly is essential, yet we recognized a gap in the evaluation of software used for this important step.

What difficulties did you run into along the way? 

Technology and software associated with genotyping-by-sequencing and de novo genome assembly are rapidly changing. During the course of our project, some of the software programs we tested were significantly updated, so we chose to rerun our analyses using the new software versions to ensure we were providing up-to-date information in our manuscript.

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

We were surprised to find such a substantial difference in performance among these assembly programs. We were especially surprised at the variation in performance among software for our first simulation where no mutations were introduced. In this scenario, we made many identical copies of genome fragments and then performed de novo assembly using each software program. Without any mutations introduced, the job is basically to generate a list of unique sequences – it should be very straightforward. In some cases, however, these genome fragments were broken into shorter sequences and rearranged beyond recognition, leading to incorrect reconstruction of the simple, unmutated data.

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

For our study, we selected a sample of assemblers from peer-reviewed literature that use different assembly algorithms, are freely available, and have updated user resources available online. However, this was not a comprehensive evaluation of all software capable of de novo assembly. Therefore, the evaluation of other programs would be valuable. Additionally, as new software programs are introduced or existing programs are updated, continued efforts to evaluate de novo assembly performance is warranted.

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

Reduced representation genotyping-by-sequencing is becoming less expensive and more accessible, making it a viable option for more research projects. While it is exciting to apply these emerging technologies and methods, it is important to recognize that approaches to filter and analyze these large datasets are still in development. Doing your background research to ensure you are applying the best available tools and using the most appropriate methods for your study is essential to doing good research in this field and in any field of research.

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

Doing this study has reaffirmed the importance of simulations to test how software works. Testing analyses on simulated data and altering parameters of the simulation or analysis can provide immense insight into how the software works and how variation in real data may affect software performance. Larger simulation projects like our study can provide information that many people can use, but I also find it incredibly helpful to run a simulated dataset through an analysis before analyzing my own data to ensure I understand what the software is doing. Taking advantage of simulated datasets available in vignettes for software is a great tool to get acquainted with the analyses you plan to do.

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

Our study demonstrates the importance of ensuring that software you use is really doing what you think it is supposed to do; and simulations can help evaluate software performance.

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

Researchers who need to perform de novo assembly of reduced representation genotyping-by-sequencing data can use our study as a guide for which software to use and the importance of different parameter settings for assembly.

LaCava, M. E., Aikens, E. O., Megna, L. C., Randolph, G., Hubbard, C., & Buerkle, C. A. (2019). Accuracy of de novo assembly of DNA sequences from double‐digest libraries varies substantially among software. Molecular ecology resources. https://doi.org/10.1111/1755-0998.13108

Interview with the authors: How does invasiveness evolve? A look at feral pigs

Understanding how and why some species readily invade new habitats is an interesting view into the myriad ways species evolve. Limiting the expansion of such introduced species can be important for managing ecosystems, particularly when the invasive species is as ecologically destructive and economically costly as the feral swine in the US south. In a paper published recently in Molecular Ecology, researchers led by Dr. Tim Smyser investigated the origins of the invasive feral swine populations to determine how much the expanding footprint of this species was a due of recently escaped domesticated pigs. Surprisingly, they found that the expanding range was largely attributable to range expansion by the established invasive swine population. Read on to for more details from Dr. Smyser into this very interesting work!

Invasive feral swine originated from a combination of European feral pigs and domesticated stock. Photo by Dr. Mirte Bosse, dvdwphotography.

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

Invasive feral swine have expanded rapidly throughout the United States over the past 30 years. The impetus for the study was to identify the drivers for that expansion, to ask: where are new feral swine populations coming from? Prior to our work, there was a hypothesis that domestic pigs had sufficient phenotypic plasticity that they would revert to a wild phenotype, resembling a wild boar, if living in the wild. Under this hypothesis, any pig farm could have served as a viable source population for invasive feral swine. With this study, we revealed that there is very little direct contribution to invasive feral swine populations from domestic pigs, potbellied pigs, or wild boar. Rather, the rapid expansion observed over the past 30 years has been driven by incremental range expansion of established invasive feral swine, which overwhelmingly represent animals of mixed European wild boar-heritage domestic breed ancestry, and long-distance translocation of feral swine from established populations to uninvaded habitats.

What difficulties did you run into along the way? 

The challenges were largely computational. We had amassed over 9,000 genotypes by the time we compiled the reference set and generated genotypes from invasive feral swine genotypes for this study. Such a large dataset required that we do everything we could to optimize runtime efficiency. Even with these efforts, the analysis still took about 4 months of runtime while using 30 CPUs with 60 threads.

What is the biggest or most surprising innovation highlighted in this study? 

I would say the most surprising result was the very high proportion of invasive feral swine that had a significant ancestry association to European wild boar. The historical record suggests wild boar releases have been far more limited than the potential for domestic pig releases, yet 97% of feral swine had significant European wild boar ancestry. This might suggest hybrid wild boar-domestic pig ancestry is biologically important for feral swine to establish self-sustaining populations and become invasive.

Moving forward, what are the next steps in this area of research?

Descending from this work, our next steps are multifaceted. With this analysis, we have identified the drivers of range expansion at a broad-scale with ancestry results pointing to the expansion of established populations. We are now interested in adding a fine-scale understanding of expansion to identify the specific sources of newly emergent populations and map the patterns of feral swine expansion. Also, this analysis has provided an understanding of the ancestral composition of invasive feral swine. Given the hybrid origin of these animals, we will identify elements of the genomes from their ancestral groups, that is heritage breeds of pig and European wild boar, that have been selectively retained in feral swine. By describing selective sweeps relative to ancestral groups, this analysis will allow us to describe the evolution of invasiveness among feral swine.   

What would your message be for students about to start developing or using novel techniques in Molecular Ecology?

The field of Molecular Ecology is changing so quickly that it is hard as a scientist to keep up, from both a computational/statistical standpoint and with all the new molecular techniques and analyses that allow us to dive deeper into the genome than we had previously imagined. My recommendation for students would be to not let the lack of a specific skill deter you from asking interesting questions – take the time to develop the needed skill sets or develop collaborations to facilitate your learning or use of those skills. Also, keep asking questions – don’t be content with the answers we are able to resolve today.

What have you learned about methods and resources development over the course of this project? 

Reflecting back on my answer immediately above, when I started asking the question of what are the drivers of invasive feral swine range expansion, I did not have the data or the skills to meaningfully address that question. Through the development of a great team of collaborators and independent learning, I was able to assemble the needed skills and then the data to pose this question and reveal interesting results. Through this project, I learned about the statistical tools used in the analyses, developed the coding skills necessary to execute those analyses, and identified strategies to maximize computational efficiency as was needed for working with such a large dataset.

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

We have demonstrated that the recent and rapid expansion of feral swine, an ecologically destructive and economically costly invasive species distributed throughout much of the US and the world, has been facilitated by movement (in many cases anthropogenic movement) from established populations to uninvaded habitats as opposed to novel introductions of either domestic pigs or wild boar.

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

In identifying the admixed origins of invasive feral swine, descending from heritage domestic pig breeds and European wild boar ancestry, we can begin to gain an understanding of the evolution of invasiveness for this species and invasive species more broadly. 

Feral Swine are not native to the U.S. They are the result of recent and historical (1500’s Spanish explorers) releases of domestic swine and Eurasian boar. USDA APHIS photo Laurie Paulik.

Smyser TJ, Tabak MA, Slootmaker C, Robeson MS, Miller RS, Bosse M, Megens H-H, Groenen MAM, Rezende Paiva S, Assis de Faria D, Blackburn HD, Schmidt BS, Piaggio AJ. 2020. Mixed ancestry from wild and domestic lineages contributes to the rapid expansion of invasive feral swine. Molecular Ecology. https://doi.org/10.1111/mec.15392

Interview with the authors: can we identify the acting selective regime in evolution experiments?

Rapid adaptation to novel conditions is an exciting and growing area in evolutionary research due, at least in part, to our desire to understand the effects of climate change, introduced species, and other conservation-related concerns. However, our ability to detect this evolution is fraught with both biological realities and technical difficulties. A recent paper by Drs. Pfenninger and Foucault, published in Molecular Ecology, illustrate how deep resequencing of replicated experimental populations can fail to provide evolutionary insights, even with extreme selective pressures, due to adaptation to unintentional environmental conditions that overwhelm the genomic signals of the intended selection. This rapid adaptation, in this case to captivity, is an interesting phenomenon that is almost certain to alter other experimental systems, including those that take place in the field. In addition to more details on this fascinating study, the interview with Dr. Pfenninger below also provides an interesting view into technical issues the research team faced: a short time after the initial publication of their manuscript, they discovered a bug in allele frequency calling software that they used! 

Swarming flight of Chironomus midges over a small puddle. Photo credit: Markus Pfenninger.

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

I wanted to know whether rapid adaptation of a natural population to an environmental stressor, in this case temperature, is possible and, if so, by which processes in detail. Apart from this being a fundamental question in population genetics, it is a crucial issue for biodiversity in the ongoing global change.

This is the official and completely true answer – but, to be completely honest, not the entire story: I wanted to see, analyse and prove evolution by natural selection hands-on. Because it’s one thing to teach something gleaned from literature and another to have seen it with your own eyes.

What difficulties did you run into along the way? 

There were actually quite a few: finding a suitable PhD student, technical difficulties with the experimental facilities that almost killed the long term experiment after some months, to mention only the most important ones.

And finally, of course, the almost detrimental issue with a bugged software tool: A few days after the official publication in January, a student reanalysing the data from a different angle, stumbled over unexplainable inconsistencies between the raw data and the allele-frequencies inferred from them. You can imagine the shock it gave me!

When I looked into the problem, I quickly found out that the allele frequencies had little to do with the raw data for most, but perfidiously not all positions in the genome, in particular not the first few on the first scaffold – that’s why the error escaped my attention during a cursory check. The allele frequencies were extracted by a software tool which indeed produced consistently wrong results – a task in principle so simple that systematic checking would have required to write a second tool that exactly did what the first should have done in the first place.

I immediately contacted the authors of the tool and they promptly confirmed that the version we used contained this bug. They did nothing wrong, though. Once they discovered the bug a few months ago, they had promptly updated the tool and documented the error in the release notes. In fact, it appears that the wrong version was on the server for a few days only. Unfortunately it was exactly during the time when we downloaded it – and who looks into the release notes after a tool seemingly did without a hitch what it was supposed to?

I had no choice but to contact the editorial office of Molecular Ecology, informing Genevieve Horn that parts of the publication were flawed and should probably be retracted. At the same time, I started reanalysing the complete data set with a correct version of the tool. Fortunately, after a hard week of number crunching, it turned out that the wrong values were highly correlated in terms of location and allele frequencies to the true values so that some numerical values, but none of the study’s conclusions, needed to be revised. The journal agreed that in this case, a correction article would be sufficient and here it is.

I have to say that everyone, from the software authors to the editor in chief, I have dealt with in this affair has responded greatly and I want to express my deep gratitude here. Given this experience with Molecular Ecology, I can only encourage everyone to address such unfortunate as perhaps unavoidable mistakes immediately and openly.

What is the biggest or most surprising innovation highlighted in this study? 

The rather unsettling major result of the study was the realisation that it is nearly impossible to experimentally manipulate the selection regime of a natural population in a targeted, predictable manner. I think, however, that such “failures” finally advance science by showing which approaches are worth pursuing and which not. Besides this more philosophical aspect showed the study the impressive power of rapid polygenic adaptation.

Moving forward, what are the next steps in this area of research?

I am currently moving into analysing population genomic time series from the field to get an idea on the selective forces acting on natural populations.

What would your message be for students about to start developing or using novel techniques in Molecular Ecology?

Have a good plan, be ready to revise it once the plan meets reality and be prepared for setbacks, remain critical about your results and incorporate appropriate controls. But perhaps most importantly, always take your time to think what you are currently doing and what should be the next steps.

What have you learned about methods and resources development over the course of this project? 

Obviously to even more thoroughly back-check every single analysis. Beyond this, I realised the value and potential of population genomic time series analysis.

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

An evolutionary experiment tells you something about the experiment – not necessarily about nature.

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

The acting selective regime in evolutionary experiments is difficult to predict and to manipulate – but perhaps it may be inferred from the results.

Pfenninger M and Foucault Q. 2020. Genomic processes underlying rapid adaptation of a natural Chironomus ripariuspopulation to unintendedly applied experimental selection pressures Molecular Ecology 59:536-548. https://doi.org/10.1111/mec.15347

Interview with the authors: Modelling multilocus selection in an individual‐based, spatially‐explicit landscape genetics framework

Genetic variation in natural systems is complex and affected by a variety of processes, and this reality has contributed to the growing popularity of simulation-based approaches that can help researchers understand the processes acting in their systems. Despite the flexibility of simulation-based approaches, simulations of natural selection across a heterogeneous landscape have typically been limited to one or two loci (e.g. Landguth, Cushman, & Johnson, 2012). In a recent issue of Molecular Ecology Resources, Landguth et al. introduce an approach to model multilocus selection in a spatially-explicit, individual-based framework, implemented in the programs CDPOP and CDMetaPOP. Read the interview with lead author Erin Landguth below to learn about the challenges in developing this program, the potential of this approach to help understand complex genotype-environment associations, and the benefits of working with strong multidisciplinary team! Read the full article here.

Dr. Erin Landguth coding in CDPOP.

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

Over the last two decades, there has been an exponential increase in landscape genetic studies, and still, the methodology and underlying theory of the field are under rapid and constant development. Furthermore, interest in simulating multilocus selection, including the ability to model more complex and realistic multivariate environmental scenarios, has been driven by the growing number of empirical genomic data sets derived from next-generation sequencing. We believe many of the major questions in landscape genetics require the development and application of sophisticated simulation tools to explore the interaction of gene flow, genetic drift, mutation, and natural selection in landscapes with a wide range of spatial and temporal complexities. Our interests lie in developing such tools and providing more flexible models that are linked to theory, and that better represent complex genetic variation in real systems. For example, adaptive traits often have a complex genetic basis that interacts with selection strength, gene flow, drift, and mutation rate in a multivariate environmental context; and this module provides the ability to simulate these processes across many adaptive and neutral loci in a landscape genetic context.

What difficulties did you run into along the way? 

When developing new modules for existing software packages, my first and primary goal is to validate these modules to theory where possible. This can take some time and many decisions, questions, and trial and errors come up along the way through this very important validation process. For multilocus selection, our validation process was to match simulation output with the theoretical expected change in allele frequencies for selection models developed by Sewall Wright in 1935. If the module is placed in the wrong location in the simulation workflow (i.e., timing) or if all of the Wright-Fisher assumptions are not matched exactly, then the simulation output will not match theoretical expectations. However, once all of these pieces are lined up, there is definitely a eureka moment, and I am then confident in the module’s performance for more complex scenarios where we will not be able to evaluate against theoretical expectations.

What is the biggest or most surprising innovation highlighted in this study? 

Multivariate environmental selection can produce complex landscape genetic patterns, even when only a few adaptive loci are involved. The relatively simple “complex” example simulated in the paper illustrates how complicated the underlying relationships can be between allele frequencies and environmental conditions. Simulating these complex relationships will be essential for testing genotype-environment association methods in a more rigorous fashion than has been seen so far. Additionally, the ability to simulate realistic landscape genetic scenarios that reflect the environmental complexity of actual landscapes will be important for validating findings from empirical data sets. 

A picture containing building

Description automatically generated
Outcome for simulation of a complex landscape and three loci. The three selection landscapes (Figure 1 of Landguth et al., 2020) are superimposed with lighter‐white areas referring to areas where all three landscapes have values of 1 and darker areas mean all three landscapes have values of −1. The copies (either 2, 1, or 0) of the first allele for each of the three loci are plotted, where darker green genotypes have more copies of these alleles (e.g., 2, 2, 2 corresponds to 2 copies of the first allele for the first, second and third loci, respectively). The first locus is associated with the categorical landscape (X1‐Figure 1a of Landguth et al., 2020). The second locus is associated with the gradient landscape (X2‐Figure 1b of Landguth et al., 2020). The third locus is associated with the habitat fragmented landscape (X3‐Figure 1c of Landguth et al., 2020).

Moving forward, what are the next steps in this area of research?

Epigenetics! We of course have a number of applications in progress for this current module, but we have already started beta testing our next module for simulating epigenetic processes in landscape genetics.

What would your message be for students about to start developing or using novel techniques in Molecular Ecology? 

Starting a simulation study in landscape genetics for the first time can be daunting and intimidating. Fear not, we say! As with all software packages, there will be a learning curve, but if you persevere and get past the first few hurdles (e.g., learning the ins and outs of file formats, running the program in a potentially unfamiliar programming interface), the door will be opened to unlimited questions that can be addressed with simulations in your system. Additionally, just like any other field study or experiment, simulation modeling is most informative when coupled with specific questions and hypotheses and well-thought-out study designs.

What have you learned about methods and resources development over the course of this project? 

As we begin to add more complex modules to these simulation platforms, I am increasingly relying on multidisciplinary approaches and teams. For example, development of this current module required Brenna Forester for her expertise in landscape ecology and genotype-by-environment concepts, as well as Andrew Eckert, with his in-depth knowledge of population genetics theory, particularly the history of additive vs. multiplicative models for fitness.

Dr. Brenna Forester, post-doctoral researcher at Colorado State University and recently awarded David H. Smith Conservation Research Fellow, helped integrate key genotype-by-environment concepts into the new module.

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

We have implemented a new module into the landscape genetic simulation programs CDPOP and CDMetaPOP that allows realistic multivariate environmental gradients to drive selection in a multilocus, individual-based, landscape genetic framework.

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

This new simulation module provides a valuable addition to the study of landscape genetics, allowing for explicit evaluation of the contributions and interactions between demography, gene flow, and selection-driven processes across multilocus genetic architectures and complex, multivariate environmental and landscape conditions.

References

Landguth EL, Forester BR, Eckert AJ, et al. (2020). Modelling multilocus selection in an individual-based, spatially-explicit landscape genetics framework. Molecular Ecology Resources, 20, 605–615. https://doi.org/10.1111/1755-0998.13121

Landguth, E. L., Cushman, S. A., & Johnson, N. A. (2012). Simulating natural selection in landscape genetics. Molecular Ecology Resources, 12, 363– 368. https://doi.org/10.1111/j.1755-0998.2011.03075.x

Wright, S. (1935). Evolution in populations in approximate equilibrium. Journal of Genetics, 30, 257– 266. https://doi.org/10.1007/BF02982240

Interview with the authors: barriers to fox gene flow in urban and rural settings

In an article published recently in the latest issue of Molecular Ecology, researchers from Researchers from the Leibniz Institute for Zoo and Wildlife Research and the Luxembourg National Museum of Natural History investigated differences between urban and rural red fox populations. They found that physical barriers in both habitats, such as a river or road, limited fox movement, and also that human activities influenced where foxes moved. This is important because it means that the interaction between human activity and other structures on the landscape may negatively alter the fox populations. For more information, please see the full article and the interview with lead author Sophia Kimmig below. 

A red fox (Vulpes vulpes) moving along rail roads in the city centre of Berlin, Germany. © Jon A. Juarez.

What led to your interest in this topic / what was the motivation for this study? Human population growth and land use are altering ecosystems worldwide and although continuing urbanization results in dramatic environmental changes, some species seem to cope well with the anthropogenic pressure. Foxes are distributed over the entire metropolitan area of Berlin, therefore it is usually assumed that they cope well with human presence. However, city life can affect key aspects of wildlife ecology and have substantial impact on the movement ecology and dispersal ability of populations. Dispersal in urban areas may be influenced by physical barriers, but also by behavioural barriers that we cannot directly see. Thus species that are physically capable of crossing the urban matrix may nevertheless face behavioural barriers due to avoidance of man-made objects (with their artificial structure, scents etc.) as well as human presence per se. We therefore wanted to understand how the landscape influences gene flow patterns in red foxes across the urban-rural matrix.

What difficulties did you run into along the way? With an increasing number of population genetic clustering approaches and R packages that differ in their precise working mechanisms, it becomes more challenging to interpret diverging results and recognize biological patterns. Further, the promising and fascinating possibilities of modelling gene flow through the landscape also come with uncertainties in how to deal with certain circumstances or type of data. For example, we discuss the issue of dealing with overlapping landscape features in the studied environment i.e. linear landscape elements (such as roads or rivers) that cross a surface structured landscape element (e.g. a forest or park). Especially in urban areas, the habitat has such a high level of complexity that you could easily spend years modelling and testing different land use layers.

What is the biggest or most surprising innovation highlighted in this study? Regarding the fox in the Berlin Metropolitan area: Foxes are quite common in urban areas, so we presumed that there would be few dispersal barriers in the urban environment. Our results have nevertheless shown that foxes disperse preferentially along linear landscape elements such as motorways and railway lines despite the inherent mortality risk. We interpreted this to mean that even urban foxes avoid the presence of humans if possible. 
Regarding a broader, biological perspective: Although we have to further improve our methods (for our study, for example, by including data on population densities, road traffic or other proxies of human presence and activity), it is fascinating that molecular genetic methods may enable us to answer more questions in behavioural ecology in the future. 

Moving forward, what are the next steps in this area of research? Now that we have familiarised ourselves with the landscape genetic techniques, we are looking forward to applying the approaches to a broad range of taxa to better understand how animals move through the landscape. This is not just of academic interest, but may help to identify and protect dispersal corridors for endangered species in a scientifically robust way.
For the Berlin foxes we are going to analyse data from a radio tracking study and research their movement patterns and space use – it will be interesting to compare those results with the ones from landscape genetics. We are looking forward to hopefully adding some more pieces to the puzzle of the city as a wildlife habitat.

What would your message be for students about to start developing or using novel techniques in Molecular Ecology? From a beginners’ perspective: For our project, we greatly benefitted from the exchange with other researchers working in this field. For example, we contacted William Peterman, who created the ResistanceGA package that we used for our landscape resistance analysis, with some methodological questions and he provided very helpful advice. I would therefore recommend getting in touch with people who work with the same methods and discussing your ideas and obstacles. Also, our work greatly benefitted from the thorough review process that it underwent. Although the requested changes and suggestions sometimes may come with a lot of re-thinking and -working effort and we usually do not always agree with every single given comment, it is crucial to take constructive criticism to improve our scientific work.

What have you learned about methods and resources development over the course of this project? Molecular genetic methods and the inherent potential to study complex ecological contexts have been changing a lot in the last decades. This is a field of frequent on-going development and improvement. Especially regarding the analytical methods, for an ecologist it is difficult to keep on track with all the latest approaches. Also, due to big data involved in the landscape analysis and the resulting time for computational analysis, the computational effort for a model becomes a real issue in landscape genetics. It is really a pity when more thorough analysis are theoretically possible and even free data is available but the analysis can just not be conducted in a feasible amount of time.

Describe the significance of this research for the general scientific community in one sentence. Assessing the impact of the habitat on (urban) wildlife beyond the physical properties of the landscape may help us to more deeply understand dispersal, behaviour and population genetic structure of populations.

Describe the significance of this research for your scientific community in one sentence. Methodological advancement due to more in depth comparisons of different genetic measures used in resistance modelling.

Kimmig ES, Behinde J, Brandt M, Schleimer A, Kramer-Schadt S, Hofer H, Börner K, Schulze C, Wittstatt U, Heddergott M, Halczok T, Staubach C, Frantz A. 2020. Beyond the landscape: resistance modeling infers physical and behavioral gene flow barriers to a mobile carnivore across a metropolitan area. Molecular Ecology. https://doi.org/10.1111/mec.15345.

Interview with the authors: historical barriers to gene flow in a fragmenting landscape

In a recent issue of Molecular Ecology, Drs. Maigret, Cox, and Weisrock published their work focused on copperhead snake response to habitat fragmentation. Interestingly, these researchers detected population structure putatively resulting from a historically important highway, even though most traffic has been shuttled to an alternative route for the last 50 years. Understanding the complexities of movement patterns in response to barriers is of increasing importance as our landscape becomes more and more fragmented. For more information, please see the full article and the interview with Dr. Maigret below. 

What led to your interest in this topic / what was the motivation for this study? The immense and rapid shift from forest to barren land and grassland which accompanies surface mining in central Appalachia is striking, especially when viewed from the air. Upwards of 20% of the land surface of some counties has been mined since 1980 through a process often termed “mountaintop removal”. The lack of research on the implications of this fragmentation was curious to me: why had such a major driver of forest loss garnered so little attention? Moreover, if we use next-generation sequencing, could we detect any effects of this land-use change on wildlife populations? It seemed like a nice natural experiment waiting to be investigated.

What difficulties did you run into along the way? Fieldwork was challenging: on top of the issues one deals with when trying to capture large numbers of secretive venomous snakes, nearly all the land in our study area is privately held, and thus gaining access to properties to collect tissue samples was time consuming. In terms of generating our data, obtaining enough DNA from our tissues (mainly scale clips) proved to be a challenge, though DNA quality was fortunately not an issue. Finally, given the diverse array of methods and subsampling protocols we used, optimizing our software pipeline took a little extra time. Thankfully, our university’s computing resources – including our associated staff and faculty – were more than adequate for the task at hand.

What is the biggest or most surprising innovation highlighted in this study? We found no evidence for an effect of mining or the current array of high-traffic roads on genetic differentiation; both of these features were hypothesized to be barriers to movement. But the most surprising part was what we did detect: a break in population similarity spatially coinciding with the path of a road which was a major highway for most of the 20th century. Previous research has suggested that highways can cleave populations of herpetofauna, and modeling work has suggested that these effects could persist for many years. We seem to have found evidence for a combination of these hypotheses, and subsampling suggested that we could have come to a similar conclusion with fewer markers and more missing data.

Moving forward, what are the next steps in this area of research? It will be interesting to see what unfolds as more genomic data is integrated into landscape genetics studies, and especially in landscapes with putative barriers of different ages or permeabilities. Re-analysis of existing data sets using (possibly) more sensitive methods, like the spatially-informed methods we used, might reveal barriers where none were detected using other approaches. As for surface coal mining, more study of the consequences of forest fragmentation – ideally, using species which might be more sensitive – could be very informative.

What would your message be for students about to start developing or using novel techniques in Molecular Ecology? Try to keep abreast of the new programs coming out. It seems like every month new approaches are being developed, and while the deluge of methods can be overwhelming at times, employing an assortment of different approaches can help enlighten one’s interpretation of genomic patterns.

What have you learned about methods and resources development over the course of this project? I’ve learned about the importance of integrating methods within an ecological framework. While a new method for analyzing genomic data is usually developed to fill a particular analytical gap, translating that goal into an ecological framework can make the method much more accessible to a broader range of researchers. And in general, doing one’s best to stay on top of the new methods coming online is important, if a little overwhelming at times.

Describe the significance of this research for the general scientific community in one sentence. Our results seem to suggest that the genomic legacy of human settlements and infrastructure can persist in wildlife populations beyond the lifespan of the infrastructure itself.

Describe the significance of this research for your scientific community in one sentence. With genomic data and statistical approaches that integrate spatial information, it might be possible to detect relatively weak genetic structuring in wild populations, and it may not require large amounts of the highest-quality data.

Maigret TA, Cox JJ, Weisrock DW. 2020. A spatial genomic approach identifies time lags and historical barriers to gene flow in a rapidly fragmenting Appalachian landscape. Molecular Ecology. https://doi.org/10.1111/mec.15362.

Interview with the authors: utilising GT‐seq for minimally invasive DNA samples

Minimally-invasive sampling is commonly used to obtain samples from rare, elusive or dangerous animals. However, this sampling technique often results in samples that are too low in quality or quantity for successful use with most high-throughput sequencing methods. Using cloacal swabs from the threatened Western Rattlesnake (Crotalus oreganus), Danielle Schmidt and colleagues show that Genotyping-in-Thousands by sequencing (GT-seq) can successfully be used to generate high-throughput sequence data from low-quality, low-quantity samples. We interviewed Danielle Schmidt (first author) and Professor Michael Russello (last author) to find out more about what went on behind-the-scenes of this study.

The Western Rattlesnake (Crotalus oreganus), a threatened species in British Columbia, Canada. Photo credit: Marcus Atkins

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

Conservation genomics has become an increasingly common term in the literature, yet many study systems that involve elusive or at-risk species must rely on minimally- or non-invasive sampling to meet research and management objectives. Although a valuable source of biological material, DNA extracted from minimally- or non-invasive samples is typically of low quantity, poor quality, and contaminated with exogenous DNA, all of which may be incompatible with modern sequencing technologies. Implementing leading-edge genetic and genomic tools to study conservation-related questions has been a long-standing interest in the Russello Lab.

What difficulties did you run into along the way?

Based on earlier work that came out of our lab (Russello et al. 2015 PeerJ), we suspected that employing a non-targeted sequencing approach like RADseq would not be efficient for collecting genotypic data from minimally-invasive samples. Therefore, we decided to test the efficacy of GT-seq (Campbell et al., 2015), as it is a targeted method that could help circumvent the typical issues involved with sequencing and genotyping lower quality DNA. Our biggest challenge was designing a GT-seq SNP panel that minimized ascertainment bias to ensure our downstream estimates of within- and among-population variation would be accurate. Also, given the number of samples and loci we planned to analyze simultaneously, optimizing the workflow for data collection took some time.

Library designs for A) RADseq and B) GT-seq. Included samples selected to facilitate within- and among-method genotype comparisons

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

One of the most surprising findings was the exceptionally high genotype consistency between paired blood and cloacal swab samples genotyped with GT-seq, and those blood samples genotyped with both RADseq and GT-seq. We even found that samples with initial concentrations as low as ~0.5 ng/uL successfully amplified, which is promising for future applications of GT-seq with minimally- and non-invasive DNA samples.

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

We are now exploring the application of GT-seq on a host of species to provide rapid, cost-effective genetic information to support research in molecular ecology and to assist wildlife and fisheries management. We are also testing the performance of this workflow with other non-invasive sample types, including feces and hair. Moving forward, we will be exploring ways of deploying these tools in the field to inform management decisions in real-time.

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

An important message we would like to convey is to think carefully about potential biases when designing a panel of markers to target, as the composition of your panel must be tailored to your research questions. For example, some applications of GT-seq may seek to intentionally maximize the among-population component of genetic variation in order to identify individuals of unknown origin to a particular fish stock with high confidence. In other cases, as with our study, we wanted a panel that could be used to most accurately reconstruct population structure and connectivity, which we were able to subsequently validate relative to a larger RADseq dataset.

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

This project highlighted the benefits of taking a new approach to address a long-standing challenge. In molecular ecology and conservation genetic studies, minimally-invasive sampling is commonly employed as either a required or a preferential approach for obtaining sufficient sample sizes. Yet, it has been recognized since the advent of non-invasive genetic sampling in the 1990’s that issues associated with DNA quality and quantity require careful consideration and extra quality control steps. Today, these considerations also apply to the use of modern DNA sequencing technologies from suboptimal starting material; however, GT-seq provides a versatile approach for overcoming DNA quality issues and providing the population-level data needed to address research and management objectives.

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

Multiplexed, amplicon DNA sequencing, such as that employed in GT-seq, is compatible with the minimally-invasive sampling often required for obtaining population-level data to inform biodiversity conservation.

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

GT‐seq offers an effective approach for genotyping minimally-invasive samples, providing accurate and precise estimates of within‐ and among‐population diversity metrics relative to genome-wide approaches such as RAD-seq.

Read the full study here:
Schmidt, Danielle A., et al. “Genotyping‐in‐Thousands by sequencing (GT‐seq) panel development and application to minimally invasive DNA samples to support studies in molecular ecology.” Molecular ecology resources (2020). https://doi.org/10.1111/1755-0998.13090

Interview with the author: Using host transcriptomics to sample blood parasites

Hosts offer diverse habitat for an incredibly rich array of microbial groups. Genomic resources for many groups residing within hosts (‘infra-communities’) are poor often due to the difficulty in isolating the DNA from the microbe from that of the host, particularly for species living within host cells. In this interview we go behind the scenes with Spencer Galen as he guides us through his transcriptomic approach he developed with colleagues to sample blood parasites such as malaria. Given how ubiquitous and important these parasites can be for animal health, this resource has the potential to pave the way for important advances in disease ecology. Read the paper here.

Avian blood transcriptomes revealed that hosts often have far more complex parasite communities than traditionally thought. For instance, the transcriptome of this Baltimore oriole (Icterus galbula) revealed at least six malaria parasite infections from three malaria parasite genera. The blood smear image from this bird shows the three genera in close contact within the host bloodstream. L: Leucocytozoon, PL: Plasmodium, PA: Parahaemoproteus.
Credit: Spencer Galen

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

This study began with two classic ingredients of scientific discovery: a lot of frustration mixed with a bit of inspiration from other researchers. The frustration was born from a lack of available genetic resources for malaria parasites and other blood parasites, which I felt was hindering the kind of research that I wanted to do. The inspiration came during the first year of my PhD, when several papers were published within a span of just a few months showing that researchers were passively generating large quantities of blood parasite genomic data by sequencing the transcriptomes of their vertebrate hosts. My PhD advisor Susan Perkins and I thought that designing a study to explore this approach in more detail could solve some of my frustrations and help the field of blood parasite research at large.

What difficulties did you run into along the way? 

When we started this project there was always the looming possibility that we would sequence a number of host transcriptomes that were infected with blood parasites and simply not recover any useful parasite data. Even a small-scale transcriptomic project is not a trivial matter financially, and so I will admit that I lost some sleep wondering if this project was a bad idea. Fortunately, field and lab work went quite smoothly, and the results of my first scan for parasites within our initial test transcriptomes exceeded my wildest expectations. And so in reality the biggest challenge was my own self-doubt – if I had paid too much attention to those thoughts, this project might not have gotten off the ground.

What is the biggest or most surprising innovation highlighted in this study? 

We were astounded by just how prevalent blood parasite transcripts can be within host transcriptomes. For instance, in one bird (Vireo plumbeus sampled in the mountains of New Mexico) we found that nearly 17% of all contigs generated from the initial Trinity assembly were derived from a parasite that was infecting just 0.75% of all blood cells. A second surprising finding was the degree to which many of the birds that we sampled were infected with complex communities of parasites that we did not detect using traditional microscopic and DNA barcoding methods. Across all samples we found that transcriptomes revealed about ~20% more infections than the methods that are typically used to study these parasites. This included one individual bird that was infected by three different genera and at least six species of malaria parasite.

Moving forward, what are the next steps in this area of research?

While it is exciting to find that a transcriptomic approach can improve our ability to study the genomic diversity and abundance of wildlife blood parasites, it still remains a rather inefficient approach – at the end of the day, the majority of transcripts from each sample came from the host organism that was not the focus of our study. The next step will be to apply single-cell and other advanced RNA sequencing techniques that have successfully been applied to model systems to provide greater resolution to studies of blood parasite gene expression and host-parasite interactions.   

What would your message be for students about to start developing or using novel techniques in Molecular Ecology? 

At risk of sounding overly pessimistic, be prepared for things to fail the first time around and have a plan B in place. It is wonderful to have a lot of confidence, but pessimism does tend to favor preparedness. Small actions within this frame of mind can save you a lot of grief in the long run, and can be as simple as testing a new method on a sample that isn’t important before you start your project or taking the time to visit a lab to learn a technique before you try it yourself. I naturally assume everything I try in the lab will fail, so each time things work (and they actually often do!) it is a pleasant surprise.

What have you learned about methods and resources development over the course of this project? 

I think that there is a difference between producing a resource, and producing a resource that is easily accessible to the broader research community in practice. As a result, I spent a lot of time thinking about how my colleagues would most directly benefit from the data that we had generated. In the end we made the data from this study available in as many formats as we thought might be useful to other researchers (raw sequences, assemblies from before and after parasite identification, curated alignments, DNA barcodes, etc.). The amount of time that it took to prepare these datasets was extremely small relative to the length of the entire project, and I think will go a long way towards making these data as useful as possible.

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

This study improves our ability to research the ecology and evolution of wildlife blood parasites, a cosmopolitan and ubiquitous group that is widely relevant to global health.

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

The methodological framework that we present in this study profoundly improves the genomic resource base that is available to research understudied blood pathogens of wildlife, as well as better detect multi-species parasite communities within hosts.