Interview with the authors: Habitat light sets the boundaries for the rapid evolution of cichlid fish vision, while sexual selection can tune it within those limits

Non-model organisms provide an interesting avenue to explore evolution in real time in natural populations. Here, we speak to Ralf F. Schneider and Sina J. Rometsch of University of Konstanz, Germany about their co-authored Molecular Ecology article, where they investigate sex‐specific opsin expression of several cichlids from Africa and the Neotropics which they coupled with data sets on sex‐specific body coloration, species‐specific visual sensitivities, lens transmission and habitat light properties. They illustrate how integrative approaches can address specific questions on the factors and mechanisms driving diversification, and the evolution of cichlid vision in particular.  Read on to get a behind-the-scenes view of this study.

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

Cichlid fishes are an amazing system to work with. They are one of the most species rich vertebrate families and adapted to a wide range of ecological niches. This is reflected in outstanding phenotypic diversity in numerous traits. Their striking variation in body coloration, which can sometimes differ considerably between the sexes, has even been acknowledged in their German name “Buntbarsche”, which translates to colorful perches. Moreover, their visual system is intriguingly complex, because cichlids possess a total of seven opsin genes that allow for color vision (humans only have three). While sexual selection has been recognized as a major driver in the evolution of cichlid body coloration (coloration being the “sender”), less is known about what shapes visual sensitivities (the receiving end of the system) in cichlids. Therefore, we were interested in whether the phenotypic diversity found across cichlid visual sensitivities is primarily driven by sexual selection (e.g., vision co-evolves with body colors), or whether environmental factors, such as light availability and water turbidity, turn out to have a stronger effect on the evolution of cichlid vision.

2. What difficulties did you run into along the way? 

The main challenge in addressing the question on whether the light environment (“ecological selection”) or conspecifics’ body colorations (“sexual selection”) are driving the diversification in cichlid vision was that both potential drivers are very complex and can be challenging to quantify. The visual environment across cichlids’ habitats varies tremendously, and is dependent on factors, such as water depth and turbidity. Coloration patterns often change across the fish’s body and only photospectrometric measurements across the whole range of visible light wavelengths can objectively quantify a color. Moreover, the fish’s visual system is highly complex. Visual sensitivity can be modified physiologically or by phenotypic plasticity by a number of factors including expressing different subsets of the seven opsin genes, changing their expression level, lens filtering etc. Combining all this information, either obtained from our own experiments or from published studies, in a common framework allowing for meaningful statistical analyses was challenging.

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

Most surprising, in terms of results, was to us that we did not find sexual dimorphism in opsin expression in any cichlid – not even those with very strong dimorphism in body coloration (such as in Pseudotropheus lombardoi (attached photo), where females are blue and males are yellow). While we did expect that evaluating mating partners based on their body color would favor associated sexual dimorphism in the visual system, this seemingly has not (yet?) happened in these fish. In terms of methodology, our study integrates a complex data-set on ecological and physiological parameters that can affect the visual sensitivity. This allowed us to evaluate the potential interactions of these parameters in a very comprehensive way.

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

In our study, we show that a wide range of data can be integrated in a single model, which allowed us to investigate interactions among variables that are rarely used in a common framework. Thus, we encourage future studies to also consider comprehensive approaches when addressing questions concerning the visual ecology of these (or other) fish, if this information is available or obtainable. Additionally, while we have a relatively good understanding of how visual information is perceived by cichlids, there is only very little information on how visual information is processed in the neuronal circuitry of the eye and later in the brain. Understanding signal processing in cichlid eyes will provide a new information layer for evolutionary ecologists to work with.

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

In the last decades, due to ever-increasing computational power, storage capacities and high-throughput techniques, such as next-gen sequencing, large amounts of data can be more reasonably collected and are accessible by more researchers. New methods can benefit from incorporating these data into analysis pipelines to consolidate them or broaden their scope. Being aware of available data can thus be very useful. However, it is also important to us to stress that approaching a scientific question from several angles and across biological disciplines that don’t frequently communicate is often the soundest approach. Classical lab methods, such as in situ hybridization or histology, as well as cutting edge techniques, such as Crispr/Cas9, can provide valuable validation/falsification of formulated hypotheses.

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

It was great to see how this study evolved: one question and technique lead to another until we finally aimed at developing an analysis frame-work for the complex data-sets that are obtained in visual ecology of (cichlid) fishes. This comprised changing and further developing our pipeline while analyzing the data. Several preliminary pipelines had to be discarded as they did not properly address our core hypotheses. Thus, an important lesson for us was that it can take a while until a newly developed analysis pipeline does actually what one envisioned roughly at the beginning of the project. Overall, collaborating in a team with members of quite different backgrounds such as ecology, molecular biology and data science and working in a large and well-established lab made it possible to learn and apply new techniques.

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

Evaluating the relative strengths of natural vs. sexual selection is a very interesting question and these two forces are often very hard to disentangle, but using a set of multidisciplinary approaches combined with a comprehensive statistical analysis allowed us to show that in narrow light environments visual sensitivity is tuned to exploit all available light, while broader light environments allow for more specialized visual sensitivities.

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

We show that ambient light is a prime driver for the evolution of visual sensitivities through natural selection in cichlid fishes, whereas sexual selection seems to finetune the observed diversity within the limits set by natural selection.

9. How has COVID-19 affected work in your group?

For the last two months we’ve all been confined to working from home which – on the plus side – allowed us to dedicate more time to data analyses and finishing manuscripts, but on the down side required lab experiments to be currently on hold – unfortunately.

Full paper: Schneider, R. F., Rometsch, S. J., Torres-Dowdall, J., & Meyer, A. (2020). Habitat light sets the boundaries for the rapid evolution of cichlid fish vision, while sexual selection can tune it within those limits. Molecular Ecology.

Interview with the authors: Strong divergent selection at multiple loci in two closely related species of ragworts adapted to high and low elevations on Mount Etna

Non-model organisms provide an interesting avenue to explore evolution in real time in natural populations.Here, we speak to Edgar Wong of Department of Plant Sciences, University of Oxford, UK about his Molecular Ecology article, which investigated speciation in two closely related Senecio species, S. aethnensis and S. chrysanthemifolius, which grow at high and low elevations, respectively, on Mount Etna, Sicily and form a hybrid zone at intermediate elevations.  Wong and his co-authors found an extremely strong selection (up to 0.78) against hybrids in the system. This estimate is one of the highest reported in literature, and much higher than the one reported in the same system in the past. Read on to get a behind-the-scenes view of this study.

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

Speciation and hybridisation have always been interesting topics to me. In the case of Senecio on Mount Etna, they have an especially fascinating story: first, Mount Etna is a relatively young mountain (less than half a million years old), and previous research hypothesized that the formation of the mountain led to the divergence of the two species, Senecio aethnensis and S. chrysanthemifolius. These species are thought to be a rare example of clear-cut, recent speciation subject to divergent selection – the formation of new species driven by adaptation to distinct conditions – high- and low-elevations in our study. Second, botanists around 300 years ago brought some live Senecio specimens of the plants from Mount Etna back to the UK, and led to hybrid speciation of S. squalidus that has since spread all over the UK (although crossing experiments using plants from Mount Etna suggested hybrid breakdown). A lot is still unknown about the plants both on Mount Etna and in the UK. Hence, I was intrigued to find out unknown aspects in the system and focused on the species on Mount Etna.

2. What difficulties did you run into along the way?

One big difficulty was that Asteraceae (which Senecio belongs to) is notorious for being hard to extract clean DNA. It was a struggle to extract good-quality DNA for this study, which was resolved in the end. Also, we only had a draft genome for the hybrid species, S. squalidus, which limited the scope of analyses we could carry out. Luckily, we managed to find some interesting, highly differentiated genes that might be underlying speciation and adaptation.

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

The most surprising finding in our study is that we estimated an extremely strong selection (up to 0.78) against hybrids in the system. This estimate is one of the highest reported in literature, and much higher than the one reported in the same system in the past. Such strong selection was surprising to us because hybrids between the two species are (apparently) happily growing at intermediate elevations between the typical habitats of ‘pure’ S. aethnensis and S. chrysanthemifolius. We think this strong cumulative selection on multiple loci works together with intrinsic incompatibility to maintain the phenotypic and genotypic divergence between the two target species.

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

In the future, we hope to identify the environmental and ecological selective forces that had shaped this system. We also hope to characterise the genetic aspect of the species by improving the genome assembly and study more in detail the intrinsic incompatibility between the two target species (such as hybrid breakdown). With more data on both extrinsic and intrinsic processes, we can integrate these findings to get a more comprehensive picture of reproduction isolation in this system.

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

I would say to spend enough time understanding the experimental techniques and different types of data analyses (and the theories behind them). Most importantly, make sure that the type of data you generate are suitable for answering your research questions. As a graduate student myself, I would also suggest not to rush your work and not get transfixed on certain issues/ problems along the way – taking a step back and asking for advice and opinions from other researchers are always helpful in getting another perspective, which often helps to find a solution.

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

The type of data I used and subsequent data analyses were all new to me when I started the project, so there is no doubt I learnt a great deal about handling new types of data and how to analyse it. Another thing I have learnt is that there are always newer or ‘better’ technologies and methods that give you more data and/ or data with higher accuracy. It is inevitable that sometimes you would be worried whether what you have is not good enough. However, I have come to realise that more isn’t always better and there will always be more advanced methods; the most important thing is to use what you have and try to answer your research questions.

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

Non-model organisms inform us a lot about evolutionary processes such as hybridisation, adaptation and speciation.

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

Strong multifarious selection could be crucial in maintaining species divergence despite on-going gene flow.

Summary from the authors: Interdependent sensory systems regulate larval settlement in a marine sponge

In the ocean, pelagic larvae that settle onto the seafloor and metamorphose into an adult directly regulate the ecology and evolution of all benthic communities. To settle, larvae of most species need to encounter specific biochemical cues that indicate an optimal environment, and many also prefer to settle in the dark. It appears likely, then, that larval responses to light and to biochemical cues are closely linked, but exactly how this happens at a molecular level is largely unexplored.

We explored how changes in gene expression regulate larval settlement in a marine sponge. We find that these larvae naturally settle at twilight, and that this is directly related to the expression of receptors and signalling pathway components. Further, we find that constant light prevents larval settlement via blocking the ability of larvae to respond to biochemical cues. Our data provide the first suggestions of candidate genes and molecular pathways that may regulate the way in which light can directly affect larval settlement. Our findings in a sponge, one of the earliest branching extant animal lineages, raises the possibility that larval responses to light and to biochemical cues might be a mechanism regulating settlement across the animal kingdom.

(Left) Scanning electron micrograph of an Amphimedon queenslandica larva. A ring of very long cilia, which are associated with photosensory pigment cells, are clearly visible at the posterior end of the larva. Photo credit: Sally Leys. (Right) Tahsha Say in the field on Heron Island Reef flat, Great Barrier Reef, Australia.

Full article: Say, TE, Degnan, SM. Molecular and behavioural evidence that interdependent photo ‐ and chemosensory systems regulate larval settlement in a marine sponge. Mol Ecol. 2020; 29: 247– 261.

This summary was written by the study’s first author,TE Say.

CRISPR-Cas Diagnostics for Environmental Monitoring

In a special blog post, Molly-Ann Williams(@WilliamsMolly_9) and Anne Parle-McDermott (@anne_parle) from the School of Biotechnology and DCU Water Institute, Dublin City University provide an overview of how CRISPR-Cas works and how it can be applied to ecology and monitoring in particular. Read their recently published Molecular Ecology Resources paper here.

The field of CRISPR-Cas for genome editing has simply exploded since its introduction in 2012. The discovery of many different Cas enzymes with additional natural or genetically engineered functionalities, is resulting in an increase in CRISPR-Cas applications across all fields from food security to medicine. 

Number of Scopus search results for query “CRISPR” in given year. Search performed on 21 November 2019 .

So how can we join the revolution and apply CRISPR-Cas to the field of Ecology?

CRISPR-Cas systems consist of two main elements: a guide and a nuclease. Guides (made of RNA) direct the nuclease (Cas enzyme) to specific nucleic acid sequences (DNA or RNA). Upon target recognition the nuclease carries out the desired response, most commonly cleavage of the target sequence. The initially discovered CRISPR-Cas system relied on a nuclease called Cas9. This enzyme is involved in highly specific cleavage of target sequences that allow genome editing to occur by activating the natural repair system of the cell. More recently the applications of this system have been expanded beyond genome editing by the discovery of several new Cas enzymes with a secondary function i.e., the indiscriminate cleavage of single stranded nucleic acids upon target recognition. The discovery of these Cas enzymes has revolutionised nucleic acid diagnostics due to two main features:

Two main elements of a CRISPR-Cas diagnostic system: Cas enzyme and guide RNA effector complex and single stranded (ss) nucleic acid reporter molecule. In this example, the nuclease is Cas12a specific to DNA detection downstream from a TTTV PAM site. Adapted from Williams MA et al (2019).
  1. Protein-guide and cleavage molecules (Cas): able to specifically recognise target nucleic acids, cleave the target sequence and subsequently cleave other non-specific nucleic acids.
  2. Nucleic acids as reporters: the non-specific nucleic acids can be designed as a reporter molecule that releases measurable signal when cleaved. This allows us to visualise when the initial target sequence has been detected and apply it to diagnostics and species monitoring.

Two main elements of a CRISPR-Cas diagnostic system: Cas enzyme and guide RNA effector complex and single stranded (ss) nucleic acid reporter molecule. In this example, the nuclease is Cas12a specific to DNA detection downstream from a TTTV PAM site.

The three main Cas enzymes of interest for diagnostics are Cas12, Cas13 and Cas14 each with unique functions applicable to different types of tests (for a more detailed discussion of these enzymes visit this blog).

The Cas enzyme most relevant for single species detection from environmental DNA is the enzyme Cas12a. This nuclease can detect both ssDNA and dsDNA but can only recognise DNA sequences downstream from a TTTV protospacer adjacent motif (PAM). Importantly, Cas12a cannot detect DNA sequences missing this PAM site. This is vital when designing single species detection assays.

Do you have two closely related species that you want to distinguish? Searching your target species sequence for a site downstream of a PAM site found ONLY in your target, and not in sympatric species, will ensure highly specific recognition and prevent detection of non-target species.

What if you work with environmental RNA? Well there is a CRISPR-Cas system for you too! The Cas enzyme Cas13 differs from Cas12a in that it recognises single stranded RNA molecules with non-specific cleavage of ssRNA following target cleavage i.e., it works the same as Cas12a but targets RNA rather than DNA.

The world of CRISPR diagnostics is still in its early stages but with the discovery of new CRISPR-Cas systems with unique functions, there is no reason ecologists cannot utilise these diagnostic tools to enhance environmental monitoring using molecular techniques. For more information on using CRISPR-Cas diagnostics for single species detection from environmental DNA read our paper here.

Methods summary: Addressing (one of) the challenges of RADseq

Article by Evan McCartney-Melstad and Brad Shaffer from University of California at Los Angeles

RADseq is a great method for gathering genomic data to answer biological questions across many different scales, from phylogenetics to population and landscape genetics. It is fast, inexpensive, and requires no previous knowledge about the species’ genomic architecture. However, with this flexibility comes challenges. In this paper we develop and bench test an approach to address what may be the biggest RADseq challenge: how to choose the right sequence similarity threshold that defines whether two non-identical sequencing reads arose from the same or different genomic locations. This problem goes to the heart of evolutionary genetics— if two sequences are considered to be homologous, or derived from the same ancestral genomic location with subsequent modification through time, then they tell us a great deal about evolutionary history. If they are paralogous, and map to separate locations, then they lack that shared evolutionary history. Getting this straight is perhaps the single most important step in using genomic data for evolutionary inference.

Heat maps showing pairwise data missingness at clustering thresholds of 88% (a) and 99% (b). 

Studies that include relatively distantly related samples, such as those asking phylogenetic or biogeographical questions, should expect that homologous sequences will have diverged over time and therefore require lower similarity thresholds that allow for that divergence. However, if the threshold is set too low, paralogs will be falsely assigned to the same genomic locus, leading to problems ranging from inflated missing data rates to inaccurate measures of genetic diversity. Rather than relying on rough guesses that are preset in software packages, our approach attempts to balance these two competing forces by quantifying the relationship between pairwise genetic relatedness (as estimated directly from the data) and summaries of the RADseq dataset including pairwise data missingness and the slope of isolation by distance among samples. The relationship between pairwise genetic distance and pairwise data missingness is particularly informative—although some positive correlation is expected as mutations accumulate in enzyme restriction sites that RAD relies on, there is often a clear pattern of increased pairwise missingness that occurs when the most divergent homologous allelic variants begin to be erroneously oversplit into different presumptive loci. By explicitly looking for this breakpoint as a function of clustering threshold, researchers can choose a value that allows them to maximize the number of genomic regions recovered while minimizing the erroneous oversplitting of highly divergent, but homologous loci.

Citation: McCartney‐Melstad, E, Gidiş, M, Shaffer, HB. An empirical pipeline for choosing the optimal clustering threshold in RADseq studies. Mol Ecol Resour. 2019; 19: 1195– 1204.

As genomic and ecological data sets grow larger in size, researchers are flooded with far more information than was available when many conventional model-based approaches were designed. To deal with these massive amounts of data, many researchers have turned to machine learning techniques, which promise the ability to help find signals within the noise of the complex data sets generated by modern sequencing approaches. Applications for machine learning in molecular ecology are broad and include global studies of biodiversity patterns, species delimitation studies, and studies of the genomic architecture of adaptation, among many others. Here at Molecular Ecology Resources, we are excited to highlight research that applies supervised and unsupervised machine learning algorithms to answer questions of interest to the readership of molecular ecology. This special issue will also highlight the nuances and limitations of machine-learning techniques. Rather than focusing on the supposed differences between machine-learning and model-based approaches, this issue would aim to highlight the broad spectrum of machine-learning approaches, many of which can incorporate model-based expectations and predictions.

We are soliciting original research that applies novel robust applications of machine learning methods on molecular data to address questions across ecological disciplines.


Manuscripts should be submitted in the usual way through the Molecular Ecology Resources website. Submissions should clearly state in the cover letter accompanying the submission that you wish the manuscript to be considered for publication as part of this special issue. Pre-submission inquiries are not necessary, but any questions can be directed to:

Special issue editors: Nick Fountain-Jones, Megan Smith & Frédéric Austerlitz

Intra-specific variation and the algal microbiome

Individuals within a species vary, and this variation can have important implications for the role a species may play within ecosystems. We compared the relative importance of variation within species due to genetic changes within its own genome versus symbiotic interactions between the focal species and its associated bacteria, also called their microbiome. We focused on Microcystis aeruginosa, a globally distributed photosynthetic cyanobacterium, also known as blue-green algae, that often dominates freshwater harmful algal blooms.

Colony of Microcystis aeruginosa from Gull Lake. Colony photographed by O. Sarnelle of Michigan State University and image prepared by John Megahan of University of Michigan.

These blooms have recently become more common and intense worldwide, causing major economic and ecological damages. We studied Microcystis and their associated microbiomes from lakes in Michigan, USA that vary in phosphorus content, which is the primary limiting nutrient in lakes. We found genomic changes among strains of Microcystis along this phosphorus gradient that indicated increased efficiency in the use of phosphorus and nitrogen. Intriguingly, we found that genotypes adapted to different nutrient environments co-occurred in phosphorus‐rich lakes. This co-occurrence may have critical implications for understanding how Microcystis blooms persist for many months, long after nutrients become depleted within lakes. Similar to previous findings in for example the human microbiome, we uncovered that the bacteria comprising the microbiomes of Microcystis varied in community composition but were more stable at the level of functional contributions to their hosts across the phosphorus gradient. Finally, while our work was mostly focused on unraveling the genomic underpinnings of nutrient adaptation, we also observed consequences of these differences in Microcystis genome and microbiome composition at a physiological level. In particular, when nutrients were provided in abundance, Microcystis (and its microbiome) that had evolved to thrive in low-phosphorus environments could not grow as rapidly as strains from high-phosphorus environments.

Sara Jackrel, Postdoctoral Fellow, University of Michigan.

Read the full article here.

Citation: Jackrel, SL, White, JD, Evans, JT, et al. Genome evolution and host‐microbiome shifts correspond with intraspecific niche divergence within harmful algal bloom‐forming Microcystis aeruginosaMol Ecol. 2019; 28: 3994– 4011.