Adaptation and drift in the deep sea - investigating the evolution of diversity in a 'uniform' environment

Lead Research Organisation: Durham University
Department Name: Biosciences


The conservation of biodiversity is necessary to provide natural populations with the potential to respond to a changing environment, and to prevent the loss of fitness associated with lost diversity. An understanding the underlying processes is necessary to promote effective transferable management strategies. This is especially important in the marine environment where there are few obvious boundaries to gene flow. Biodiversity is lost within populations by drift & directional selection, and retained among populations by the same processes. The size of insular populations helps to determine the dominant process, since drift is stronger than selection in small populations. It is known that much natural diversity is divided among conspecific populations by these processes, but the fragmented structure of this diversity is often cryptic, especially in poorly understood environments like the deep sea.

We typically define populations a priori by geographic distance and according to apparent physical boundaries. However, in addition to spatial factors, environmental characteristics can affect patterns of connectivity by affecting the rate and direction of gene flow, and this influence will vary as environments change over time. Environmental characteristics also drive local selection, and selection can maintain differences among populations even with continuing gene flow. Without knowledge about how local populations may have adapted to local environments, we remain unable to incorporate that type of biodiversity into conservation strategy, and to address its potential importance to the long-term survival of species. Understanding the function and value of biodiversity requires a better understanding of the interacting processes of selection and drift leading to its evolution and loss. While much of the relevant theory is well established, empirical data are required to test theory and determine how environmental factors interact with evolutionary processes in natural systems.

In this study we will employ second generation sequencing technologies to investigate how environmental factors are associated with the evolution of population structure and speciation in the deep sea, at both neutral and functional loci. The particular focus will be on habitat depth and isolation by distance, however other environmental factors will be quantified at collection sites based on data collected from the same sites during the recent ECOMAR consortium study (during which samples for the proposed study were collected). The study species will be fish in the genus Coryphaenoides, a specious genus of deep demersal fishes, some of which are commercially important fisheries species (including the primary focal species in this study, C. rupestris). Initial studies have shown phylogenetic lineage division correleted to habitat depth. Other work has indicated selection at genes associated with adaptation to pressure in some of these species. Two focal Coryphaenoides species will be investigated in a comparative population genomics study. These species, C. rupestris and C. brevibarbis are sympatric, but the former is found at shallower depths than the latter. We will test hypotheses about how habitat division by depth or other characteristics (e.g. current systems or geographic distance) may determine the evolution of intra- and inter-specific diversity (including among multiple habitat-specialist species in the genus using an extended phylogenetic assessment) both by drift (through differential environmental pressures associated with dispersal) and by selection. Greater connectivity is predicted in the abyssal habitat, which may lead to larger effective population sizes and facilitate adaptive evolution. Taken together these data will provide novel insight into the interactive role of drift and selection during the evolution of diversity, and transferable inference about how diversity is structured in the deep sea.

Planned Impact

The primary impact of this work beyond the beneficiaries in the academic community will be on fisheries management and commercial fisheries working in the deep sea. In the eastern North Atlantic (including the high seas area out to and including the mid-Atlantic ridge) deep-sea fisheries management comes under the jurisdiction of the European Commission and the North-East Fisheries Commission (NEAFC). Deep-sea stocks are defined as those living at depths of 400m or greater. Although fishing by factory trawlers and modern longliners started in the eastern North Atlantic in the late 1960s, regulation was introduced only recently, beginning in 2002 with the introduction of the first 'total allowable catch' (TAC) quotas (setting quotas for 2003). This covered eight species, among which was the roundnose grenadier (Coryphaenoides rupestris, a focal subject species in this study). Two further species were later added together with a quota for a collection of 11 deep-sea shark species (considered together) for 2005 and 2006. TAC allowances are now set for a given EC country and ICES fisheries region separately. One recognised problem with this approach is that deep-sea fisheries are typically multi-species in character, based on trawling or gil-net fisheries. A targeted species allowance assumes that there is a simple relationship between the recorded landing records and effort for that species, which is unlikely to be the case in multi-species fisheries. This could mean that there is greater impact from fisheries for a given species than indicated in the catch records, and so greater impact on genetic diversity. In 2009 the NEAFC agreed to adopt a measure that closed 330,000 square km to bottom fishing at the mid-Atlantic ridge, including our core study area in that region near the Charlie Gibbs Fracture Zone. There are concerns about the effectiveness of this measure as well, as many of the target species have different geographic distributions at different life stages, likely including the grenadier species in this study, due to larval drift on oceanic currents (and regions near the study area at the mid-Atlantic ridge remain a focus of roundnose grenadier fisheries).

Our data will help to resolve this question. The roundnose grenadier has had the highest quota assigned among all of the managed species, with a total of 12,443 tonnes allowed in 2005 falling to 8,109 tonnes for 2011 (but still the highest TAC quota for a deep-sea fish). However, little is know for any of the subject species about whether or not regional populations should be managed separately. Our study will address this question for the roundnose grenadier for most regions of primary importance to the fishery (ICES areas VI, VII & XII) using a very high resolution methodology, however of greater importance will be the general assessment of drivers of stock structuring in the deep sea, and insight into the best methods for assigning management areas. We will also provide detailed information on the distribution of diversity at functional genetic markers for the first time for a deep sea fish species, and assess their potential relevance in biodiversity conservation and management. Details of our findings will be brought to the attention of the NEAFC and European Commission for consideration. Deep-sea fisheries managers will be among those included in our program of end-user engagement. Other interested end users will include the general public with an interest in life in the deep sea. Through recent involvement in the Census of Marine Life project MARECO (focussed on life in the deep sea at the mid-Atlantic ridge), it was evident that our outreach for that project was highly popular and successful, indicating a broad interest in discoveries about this largely unknown ecosystem. We will establish a web page providing explanations for a lay audience, updates, images and news and also publish in the popular media.


10 25 50
Description We are investigating processes across a broad evolutionary scale from the forces driving genetic structure among populations to the processes driving the generation of new taxa, and especially in the context of adaptation to living in the deep oceans. For this work we are focusing on the deep-sea fishes in the genus Coryphaenoides, of which there are 65 recognized species. To assess the evolutionary relationships among species and test hypotheses concerning drivers of divergence we have produced the most complete phylogeny for the group to date. Using 3,048 bp of Sanger sequence data across 28 species, our analyses show divisions among species that correlate with depth. We have found that species that are able to reside at abyssal depths form a distinct and well supported lineage. The finding that all abyssal species fall into a single lineage indicates that colonization of the abyss happened just once and that movement into the deepest depths of the oceans requires special adaptation. Our manuscript on this topic was published in Molecular Phylogenetics and Evolution in 2016.

To investigate the processes driving population subdivision (within species) we have produced over 800 billion bases (GB) of genomic 'RADseq' data across approximately 400 individuals in the two focal species in this study (C. brevibarbis and C. rupestris). Using these data to improve our understanding of the way that key environmental factors and geographic features influence population structure and local adaptation we developed >7000 single nucleotide polymorphic (SNP) loci across 5 populations of C. brevibarbis along the Mid-Atlantic Ridge. Our data matrix of 102 individuals is now complete. Analyses of the neutral loci indicate only weak population structure across the region (detected only using ordination methods). In contrast, signals of population structure in outlier loci indicate that selection may be acting on populations particularly across the Charlie Gibbs Fracture Zone. We decided to incorporate our C. brevibarbis genomic data (see below) into the analyses, and are now finalizing this work and are preparing the manuscript. The C. rupestris dataset includes approximately 300 individuals across 9 populations in the North Atlantic including one population in Skagerak, Norway. We have now completed RADseq analyses both for the 9 regional populations and for deep and shallow water samples around the Hebrides considered separately. For the former we used 4989 neutral SNPs and for the latter 6657. We have mapped the outlier reads against our C. rupestris genome to improve resolution and to identify outlier loci, potentially under selection. These data reinforce the findings associated with differential selection at different depths. We completed resequencing the genomes (>5x coverage) of 60 C. rupestris from a single location across a depth gradient (750, 1000, 1500, and 1800m). These genomes were mapped back to the genome assembly (described below) to identify millions of loci for population level analyses. Those analyses identified loci associated with habitat depth with fixed genetic differences within coding genes at non-synonymous sites. Roughly 30 loci showed strong outlier status, based on the comparison by depth, using 'Manhattan' plot comparisons. We have genotyped individuals at some of these loci more widely, and analysed these data in the context of Hardy-Weinberg expectations. We find that disruptive selection based on habitat depth is supported across the species range in the North Pacific. Combining data about life history, ecology, and the genotypes of juveniles, our paper on this aspect of the study reveals novel inference about sympatric ecotype evolution and the role of habitat depth in shaping adaptive evolution in the deep sea - published in Nature Ecology and Evolution (2018, 2, 680-687 doi:

We have learned novel inference about stock structure within and among geographic regions, and tested hypotheses about the implications for the role of life history and environmental factors (especially depth) in the evolution and loss of diversity in this ecologically important species. To more fully investigate the role of depth in driving divergence in deep-sea organisms we sequenced the genomes of 14 species across the genus including 7 species from the abyssal lineage. A high coverage assembly of the C. rupestris genome (0.83GB at 120x coverage - now online at NCBI and described in the Nature Ecology & Evolution paper) has been achieved via a combination of PacBio and Illumina sequencing. We also produced RNAseq data for different tissue types from C. rupestis which was used to facilitate annotation of the C. rupestris genome, providing a high-quality genome for this species with annotation to a good level of resolution. Additionally, a 91x coverage assembly of the C. brevibarbis genome and 5x coverage of the other 12 genomes have been produced and mapped back to the C. rupestris assembly. Further refinement of the C. brevibarbis genome has been achieved using DoveTail methodologies. The goal of this ambitious effort is twofold: to produce annotated genomes of deep-sea fishes that will be invaluable resources to the broader scientific community and to identify regions of the genome that correlate with adaptation to depth. By directly comparing the genomes of these species we will identify regions that may have been involved in adaptation to abyssal depths. Most of the bioinformatic work has been completed for this aspect of the project, and initial drafts of the papers have been outlined. As a further extension of this work we are sequencing loci showing evidence for strong selection associated with habitat depth for a broader range of teleost fish found at different depths. This aspect of the project is still underway.

Another on-going study building on the earlier work is comparing a new set of 290 fish sampled in the earlier transect area for age and genotype at the non-synonymous outlier loci to test the hypothesis that selection is on-going among cohorts. Genotyping was done in collaboration with colleagues at the University of California, Santa Cruz using the 'genotyping by thousands' technique. Ageing was completed in Durham by interns together with collaborators from Marine Scotland. While these data are still being analysed, the preliminary results are very interesting, showing a relationship between the proportion of genotypes and age, deviating from Hardy Weinberg equilibrium only in the older fish. We would like to increase sample sizes per age class, and so have determined the age of an additional 139 fish from across the depth range, and are planning to genotype these new individuals soon. In this part of the study we will test the hypothesis that ongoing selection is preserving ecotype diversity within a panmictic population.

A further development was initiated in 2018 and will be completed this year, supported by a Master's student working on morphometric analyses and protein modelling to further investigate the adaptive differences among fish living at different depths. This study is also in collaboration with colleagues at Marine Scotland. There were a number of significant differences in morphology among fish collected at different depths. For example, fish from the deeper environment had smaller adult body size, a more elongated body form, broader mouth gape, and a series of other morphological distinctions. A random forest analysis could easily classify them to the correct depth based on morphology. However, as shown in our Nature Ecology & Evolution paper, fish from different depths are not differentiated at neutral genetic markers (based on comparisons at up to millions of SNP loci). This suggests either that different phenotypes that live at ~1000m as juveniles, migrate to different depths as adults, or that the change in morphology reflects developmental plasticity. Our results to date would favour the former explanation. For our protein structure analyses we have focused on muscle structural proteins such as obscurin (OBSL1), and found functional changes that are consistent with our broader interpretation about adaptation to habitat.

As we continue to expand on these analyses and generate new data, we find it appropriate to delay publications until we develop a more complete understanding of this system. However it is expected that multiple papers will be ready at around the same time, and significantly increase the level of output from this project.
Exploitation Route To ensure that the results of our work reach end-users we have multiple further manuscripts in prep in addition to papers already published. Also, in year 1 we put together a panel of research biologists and resource managers to provide feedback and make suggestions concerning our work. Each is an expert in different aspects of deep-sea biology, fisheries, and conservation. The working group met via virtual workshops and resulted in further collaborations for broader impact. Working Group members 1) Monty Priede, Professor Emeritus at University of Aberdeen 2) Rachel Jeffreys, Royal Netherlands Institute for Sea Research 3) Alex Rogers, Professor in Conservation Biology at the University of Oxford. 4) Pascal Lorance, Marine Biologist at IFREMER 5) John Gordon, Research Fellow at the Scottish Association of Marine Science 6) Kerry Howell, Associate Professor at Plymouth University 7) Francis Neat, Senior Researcher at Marine Scotland Science 8) Ross Jolliffe, Divisional fisheries Director at Centre for Environment, Fisheries and Aquaculture Science (Cefas) 9) Tom Blasdale, Marine Species Adviser at the Joint Nature Conservation Committee (JNCC) 10) Oscar Gaggiotti, Professor at the University of St Andrews 11) Selina Stead, Professor of Marine Governance and Environmental Science at Newcastle University 12) Ben Wigham, Lecturer in Marine Science at Newcastle University. We will emphasise results from our study, working with stake holders and policy makers, to help ensure improved fisheries management strategies.

We also constructed a website ( to help disseminate information both to the public and to relevant conservation agencies. The website has a front end that explains the science in lay terms, describes the problems faced by resource managers, and outlines how our project will address management concerns. The website includes a blog for public feedback, summary of working group meetings and links to other relevant and important websites.

All data will be made available on public databases, and will provide a significant and novel resource, including high quality genomes for two species living in environments that generate selective pressures associated with physiological and behavioural adaptation to depth. Extensive draft genomic data will be made available for an additional 12 species in the same genus. Our data on specific loci that are associated with living at different ocean depths will provide raw material for those interested in protein function and adaptive change in the context of environmental change. Our paper in Nature Ecology and Evolution has attracted attention in the media, on twitter and via blogs.
Sectors Agriculture, Food and Drink,Education,Environment

Description While we are still analyzing data, we have manuscripts in production and more to come in the near future. The PI and PDRA on the project have given numerous talks about the project and its potential impact and we have been disseminating ideas and engaged in discussion through our working group (a panel of research biologists and resource managers established to provide feedback, make suggestions, and ask questions - each is an expert in different aspects of deep-sea biology, fisheries, and conservation). Our final meeting with the working group was very productive, and two interactions have continued, facilitating further inference and interpretation. In particular, we are now developing a stronger link between work on the ecology and conservation management of the subject species, studied at Marine Scotland, and our data from genomic analyses. We are finding increasingly compelling evidence for the need to manage population more carefully by habitat depth, and a major publication on the subject is in preparation, including collaborative analyses with colleagues at Marine Scotland. We have maintained our webpage which is still generating interest as indicated by traffic on the blog.
First Year Of Impact 2014
Sector Communities and Social Services/Policy,Education,Environment
Impact Types Societal

Description Adaptation of Coryphaenoides rupestris along a depth gradient 
Organisation Marine Scotland Science (MSS)
Country United Kingdom 
Sector Public 
PI Contribution We are extending our analyses of adaptation along a depth gradient for Coryphaenoides rupestris by analysing ~350 further samples, aging individual fish, genotyping at loci identified under selection in earlier study, undertaking morphometric analyses, and modelling protein structure. The work is being undertaken by a student intern and a Master's degree student.
Collaborator Contribution At the Southwest Fisheries Science Center, Santa Cruz, CA, USA samples are being genotyped for key loci known to be under selection. At marine Scotland Science additional fish are being collected, replicate ageing of otoliths is being undertaken, and collaborative statistical analyses are being completed.
Impact Work still underway, but at an advanced stage. Work incorporates, population genetic, ecological, life history and evolutionary analyses.
Start Year 2016
Description Adaptation of Coryphaenoides rupestris along a depth gradient 
Organisation National Oceanic And Atmospheric Administration
Department Southwest Fisheries Science Center
Country United States 
Sector Public 
PI Contribution We are extending our analyses of adaptation along a depth gradient for Coryphaenoides rupestris by analysing ~350 further samples, aging individual fish, genotyping at loci identified under selection in earlier study, undertaking morphometric analyses, and modelling protein structure. The work is being undertaken by a student intern and a Master's degree student.
Collaborator Contribution At the Southwest Fisheries Science Center, Santa Cruz, CA, USA samples are being genotyped for key loci known to be under selection. At marine Scotland Science additional fish are being collected, replicate ageing of otoliths is being undertaken, and collaborative statistical analyses are being completed.
Impact Work still underway, but at an advanced stage. Work incorporates, population genetic, ecological, life history and evolutionary analyses.
Start Year 2016
Description Field logistics and samples 
Organisation Marine Scotland, Fisheries Research Services
Country United Kingdom 
Sector Public 
PI Contribution We will use material provided to us (and in the case of Marine Scotland, the boat time provided) to further or genomic research and we will include co-authors from the partner organisations on our publications. We will also provide important information to facilitate the more effective conservation and management of the study species.
Collaborator Contribution This entry covers a broad range of contributions, always involving the provision of species materials in support of our research. Boat time and space for our project PDRA on board was provided by Marine Scotland - all other collaborations involved shipping available samples. The list of collaborators is as follows: Jeffery Drazen, University of Hawaii Dean Grubbs, Florida State University Adela Roa-Veron, VIMS Tracey Sutton, NOVA Alastair Graham, CSIRO Ingvar Byrkjedal, Bergen Museum Guldborg Sovik, Institute of Marine Science, Norway Odd Aksel Bergstad, Institute of Marine Science, Norway Andrew Bentley, Kansas University Biodiversity Institute Takashi P. Satoh, National Museum of Nature and Science, Japan Katherine P. Maslenikov, The University of Washington Fish Collection Jerry Hoff, NOAA Daniel J. Kamikawa, NOAA Nancy Roberson, NOAA Finlay Burns, Marine Scotland Francis Neat, Marine Scotland
Impact Papers are still in preparation
Start Year 2013
Description Genomics 
Organisation University of Liverpool
Department Liverpool Centre for Genomic Research
Country United Kingdom 
Sector Academic/University 
PI Contribution Genomic work on this project was done in collaboration with Prof. Neal Hall and the Genomics Centre at Liverpool University.
Collaborator Contribution RAD sequencing was undertaken at the Genomics Centre in Liverpool for project NE/J014443/1. For project NE/K005359/1 the collaboration is ongoing, and genome sequencing, re-sequencing, genome annotation and associated bioinformatics work is being done in Liverpool.
Impact The data generated for NE/J014443/1 on population genomics and phylogenomics (see publications) were an outcome of this collaboration. Outputs for the current project are currently in preparation.
Start Year 2012