Fisheries induced evolution in Atlantic cod investigated by ancient and historic samples

We focus on the reconstruction of whole genome sequence (WGS) data from historic populations of Atlantic cod, and compare these to their contemporary counterparts. The historic populations come either from before the beginning or during the 20th century, and includes the onset of large-scale industrial fishing. Hence this comparison aims to quantify the selective effects that industrial fishing has wrought upon Atlantic cod. During this project, we had the opportunity to obtain a few Atlantic cod samples from archaeological excavations dated to the Viking Age. Since these samples have the potential to provide a unique long-term perspective, we included a limited number of these ancient samples into this project. Using such samples, we may be able to reconstruct of past-trade routes, and to improve our understanding of the earliest impacts of large-scale fishing. Overall, we analyze these ancient, historic and contemporary populations pairs from Canada, Iceland and Norway, characterizing temporal changes over a large part of natural range of Atlantic cod. We have obtained WGS data from hundreds of contemporary specimens (n > 600, also in collaboration with the Aqua Genome Project), historical specimens (n > 160) and ancient specimens (n = 9). Contemporary and historical samples have been sequenced to approximate 8X coverage, the ancient specimens to approximately 2X. It is well known that data from historic samples may be biased due to ancient DNA degradation. Nonetheless, during our work we encountered two novel types of bias that were not observed before. First, a popular library preparation method may lead to a specific error whereby the beginning of an ancient DNA fragment is erroneously copied towards the end of that same fragment (http://dx.doi.org/10.1371/journal.pone.0089676). Second, we find that historic DNA data from Atlantic cod can consist of 50% simple repetitive DNA of a specific kind - a so-called dinucleotide repeat. While all vertebrates have this type of repeat, it is particularly prevalent in the Atlantic cod genome (see also http://dx.doi.org/10.1101/060921). We found that - under certain conditions - these repeats may be preferentially amplified, leading to an overrepresentation in the dataset (http://dx.doi.org/10.1080/20548923.2016.1160594). Both types of bias can severely affect sequencing efficiency and may influence down stream analysis if not addressed. Since most aDNA extraction protocols are designed for heavy mammalian bones or teeth, we developed a protocol to deal with the brittle, porous and flimsy fish bones. This method is particularly successful and shows that ancient fish bones can be an excellent source of DNA (doi: 10.1111/1755-0998.12623). We wished to be able to determine genetic sex in historical samples. For this, we used contemporary material to map the sex-locus of Atlantic cod to a small genetic region (doi:10.1038/srep31235). Knowledge of the sex-locus allows us to investigate sex bias in historical and ancient samples. To our knowledge, we have now assembled the world's largest genomic dataset on a spatiotemporal scale for a major marine resource. Most striking is the discovery of three large chromosomal rearrangements in these populations. These rearrangements strongly contribute to contemporary population structure, which underscores the importance of these regions for local adaptation. Indeed, 98% of the genomic variation that substantially contributes to population divergence is located in just 5% of the genome. Intriguingly, while most of the genome shows temporal stability (i.e. no change over time), one region does show putative patterns of change although sampling bias or demographic change may confound such conclusion. Overall, preliminary evidence suggests that industrial fishing did not affect large effect loci over the last 120 years. More detailed analysis to look for epistasis among loci and the existence of many loci of small effect are ongoing. The available data is so abundant that we have hired a PhD candidate through the MARMAED Innovative Training Network (http://www.marmaed.uio.no/) funded through the Horizons 2020 initiative and headed by CEES. This PhD will perform time-series analysis using this dataset for the next 2 years. We have further initiated collaborations with Dr. Malin Pinsky from Rutgers University, USA to work on these data. Finally we are analyzing the sample origin of nine ancient fish genomes from five archaeological excavations and find that Atlantic cod from Lofoten may be found in early Viking Age (700-1000AD) locations as far as Schleswig or Hedeby (Haithabu) in Germany. Since the earliest evidence so far suggested that this trade did not develop until after 1100AD, these data indicate that the demands of long-distance trade may have impacted this marine resource for a more extended period than expected.

The fisheries on Atlantic cod (Gadus morhua) can be considered archetypal, with a long history of exploitation and prolific examples of stock depletion and collapse. Concern about the sustainability of this and other fisheries has stimulated research to i nvestigate the extent of human impact on marine ecosystems. One of these impacts, fisheries induced evolution (FIE) has long been suspected in driving long-term changes in phenotypic traits of heavily exploited fish stocks around the world. These changes are expected to be detrimental, for example by lowering growth-rate or age-at-maturation, and may lead to substantial economical damage or, even more dramatically, lead to the complete collapse of individual stocks. Direct empirical support for FIE, howev er, has been lacking and instead the evidence has been indirect, implicated through phenotypic trends observed in major commercial fisheries. One reason for this lack of evidence is the absence of advanced genomic tools specifically designed to target tho se species subjected to FIE. Here we will investigate direct evidence for FIE and assess its impact, by exploiting recent advances in genomics, in particular the completion of an annotated reference genome for cod. We will compare the genomes obtained fro m historic cod populations that have experienced distinctive selective scenarios, to those of their contemporary counterparts. The temporal characterization of genome-wide variation in these populations allows the statistical evaluation the selective pote ntial of novel genomic regions and candidate genes. Using the availability of archived and archeological material, we will further scrutinize a subset of most promising genomic regions. By evaluating these regions on a temporal scale, and associating chan ges in genomic variation with distinct selective scenarios over time, we will be able to further disentangle the impact of FIE and climate change in this important ecological and economical species.