Oceans constitute over 70% of the earth's surface, and the marine environment and ecosystems are central to many global challenges. Not only are the oceans an important source of food and other resources, but they also play a important roles in the earth's climate and provide crucial ecosystem services. To monitor the environment and ensure sustainable exploitation of marine resources, extensive data collection and analysis efforts form the backbone of management programmes on global, regional, or national levels. Technological advances in sensor technology, autonomous platforms, and information and communications technology now allow marine scientists to collect data in larger volumes than ever before. But our capacity for data analysis has not progressed comparably, and the growing discrepancy is becoming a major bottleneck for effective use of the available data, as well as an obstacle to scaling up data collection further. Recent years have seen rapid advances in the fields of artificial intelligence and machine learning, and in particular, so-called deep learning systems are now able to solve complex tasks that previously required human expertise. This technology is directly applicable to many important data analysis problems and it will provide tools that are needed to solve many complex challenges in marine science and resource management. Here we give a brief review of recent developments in deep learning, and highlight the many opportunities and challenges for effective adoption of this technology across the marine sciences.

Motivation: The commercial launch of 454 pyrosequencing in 2005 was a milestone in genome sequencing in terms of performance and cost. Throughout the three available releases, average read lengths have increased to ∼500 base pairs and are thus approaching read lengths obtained from traditional Sanger sequencing. Study design of sequencing projects would benefit from being able to simulate experiments. Results: We explore 454 raw data to investigate its characteristics and derive empirical distributions for the flow values generated by pyrosequencing. Based on our findings, we implement Flowsim, a simulator that generates realistic pyrosequencing data files of arbitrary size from a given set of input DNA sequences. We finally use our simulator to examine the impact of sequence lengths on the results of concrete whole-genome assemblies, and we suggest its use in planning of sequencing projects, benchmarking of assembly methods and other fields. Availability: Flowsim is freely available under the General Public License from http://blog.malde.org/index.php/flowsim/ Contact: susanne.balzer@imr.no; ketil.malde@imr.no

Calcium (Ca) - fortified foods are likely to play an important role in helping the consumer achieve an adequate Ca intake, especially for persons with a low intake of dairy products. Fish bones have a high Ca content, and huge quantities of this raw material are available as a by-product from the fish industry. Previously, emphasis has been on producing high quality products from fish by-products by use of bacterial proteases. However, documentation of the nutritional value of the enzymatically rinsed Ca-rich bone fraction remains unexplored. The objective of the present study was to assess the bioavailability of calcium in bones of Atlantic salmon (oily fish) and Atlantic cod (lean fish) in a double-blinded randomised crossover design.

The salmon louse, Lepeophtheirus salmonis, is an ectoparasite of salmonids that causes huge economic losses in salmon farming, and has also been causatively linked with declines of wild salmonid populations. Lice control on farms is reliant upon a few groups of pesticides that have all shown time-limited efficiency due to resistance development. However, to date, this example of human-induced evolution is poorly documented at the population level due to the lack of molecular tools. As such, important evolutionary and management questions, linked to the development and dispersal of pesticide resistance in this parasite, remain unanswered. Here, we introduce the first Single Nucleotide Polymorphism (SNP) array for the salmon louse, which includes 6000 markers, and present a population genomic scan using this array on 576 lice from twelve farms distributed across the North Atlantic. Our results support the hypothesis of a single panmictic population of lice in the Atlantic, and importantly, revealed very strong selective sweeps on linkage groups 1 and 5. These sweeps included candidate genes potentially connected to pesticide resistance. After genotyping a further 576 lice from 12 full sibling families, a genome-wide association analysis established a highly significant association between the major sweep on linkage group 5 and resistance to emamectin benzoate, the most widely used pesticide in salmonid aquaculture for more than a decade. The analysis of conserved haplotypes across samples from the Atlantic strongly suggests that emamectin benzoate resistance developed at a single source, and rapidly spread across the Atlantic within the period 1999 when the chemical was first introduced, to 2010 when samples for the present study were obtained. These results provide unique insights into the development and spread of pesticide resistance in the marine environment, and identify a small genomic region strongly linked to emamectin benzoate resistance. Finally, these results have highly significant implications for the way pesticide resistance is considered and managed within the aquaculture industry.

The age structure of a fish population has important implications for recruitment processes and population fluctuations, and is a key input to fisheries-assessment models. The current method of determining age structure relies on manually reading age from otoliths, and the process is labor intensive and dependent on specialist expertise. Recent advances in machine learning have provided methods that have been remarkably successful in a variety of settings, with potential to automate analysis that previously required manual curation. Machine learning models have previously been successfully applied to object recognition and similar image analysis tasks. Here we investigate whether deep learning models can also be used for estimating the age of otoliths from images. We adapt a pre-trained convolutional neural network designed for object recognition, to estimate the age of fish from otolith images. The model is trained and validated on a large collection of images of Greenland halibut otoliths. We show that the model works well, and that its precision is comparable to documented precision obtained by human experts. Automating this analysis may help to improve consistency, lower cost, and increase the extent of age estimation. Given that adequate data are available, this method could also be used to estimate age of other species using images of otoliths or fish scales.

Motivation: The nucleotide sequencing process produces not only the sequence of nucleotides, but also associated quality values. Quality values provide valuable information, but are primarily used only for trimming sequences and generally ignored in subsequent analyses.