Tracing phytoplankton grazed by mussels using molecular methods to identify preys and improve modelling

Advances in ecological modelling have motivated the development of operational models to support management of environment and resources. Still, many key processes in ecological models are poorly represented and resolved; either because we do not understand the mechanisms or because we are not able to provide data to parameterize, force and test models beyond certain levels of detail and complexity. Feeding is one such key processes that regulate mass and energy transfer between organisms and trophic levels, and ecosystem models cannot deliver reliable predictions unless the data input on these two processes have adequate resolution. In most predictive models today, suspension feeding is often reduced to a simple function of food concentration of homogeneous food items, which often fail to reflect empirical observations, hence the predictions become unreliable. Here we use the mussel Mytilus edulis as a model species, to improve knowledge on feeding processes in suspension feeders to parameterize the variations inherent in the Dynamic Energy Budget (DEB) growth model. The DEB model is a generic model applicable to several marine species, and it has potential application in ecosystem modeling systems. Results from conventional feeding experiments using high-resolution particle detection combined with molecular tools tracing phytoplankton shows how the mussel select on size and between phytoplankton from a natural seston environment. The experiments were conducted in the Lysefjord (western Norway), a fjord periodically enhanced by controlled upwelling of nutrient rich deeper water stimulating primary production. The project conducted in depth analysis of potential methodological influences on the determination of particle retention efficiency (RE) of mussels. RE dependence on particle size has previously been described as a marked decline below maximum retention of particles larger than 3-7 µm. For all particle suspensions and methods, mussels in Lysefjord had a maximum RE for particles greater than 8-11 µm. The RE for smaller particles declined gradually with 50-60 % retention of 4-µm particles and 30-40 % retention of 2-µm particles. The results showed that mussel feeding does not always conform to the traditionally assumed model, and contradicts the widely accepted beliefs in the field of bivalve ecophysiology. This has contributed to many conclusions on their ecosystem interactions and has been applied in formulation of feeding in DEB modelling. The project recommends on methodology of clearance rate in order to get reliable measurements. This is crucial in assessing the parameters representing feeding in DEB modelling. To understand any possible selection independent of particle size, molecular trophic interaction techniques were used to compare DNA signatures from seston available and ingested by mussels. Two general approaches were used in order to provide added resolution on ingestion of seston; 1) a metabarcoding approach targeting the V7 region of the 18S rRNA gene and sequenced by the Ion Torrent PGM platform enabling an overview of the complete seston community, and 2) a quantitative PCR approach targeting spiked algae added to the seston community where each algae were traced from the inlet to outlet siphon and into the digestive system. Comparison of the seston composition in the natural water and digestive system suggests that prey could be classified into three groups: A first group of preys, which were more abundant in the mussel digestive system than in the inlet water, suggesting selection process. A second, with equal abundance in both inlet water and the digestive system, suggesting no selection. Finally, a third with higher abundance in the natural water than in the digestive system, suggesting rejection of these prey particles. Data from the spiking experiment support the results from the metabarcoding analysis were larger and chain-forming species was preferentially retained. To our knowledge, the present project represents the first use of metabarcoding on mussel diet analysis. Since mussels in the Lysefjord show high growth related to reproductive activity (subsequent to spawning) specific method for parameterization in DEB has been tested with focus on the parameters determining feeding and partitioning energy between structure and reproduction. Results from the high-resolution measurements and molecular studies of feeding provide new information in establishing a mechanistic description of the suspension feeding process in DEB, with potential application to ecosystem modelling systems. The project results contribute to fundamental understanding of bivalve feeding influencing the regulation of mass and energy transfer between organisms and trophic levels. This adjust our knowledge on how suspension feeders interact with the environment, and their role in processes important for industry and society.

Advances in ecological modelling have motivated the development of operational models to support management of environment and resources. Still, many key processes in ecological models are poorly represented and resolved; either because we do not understa nd the mechanisms or because we are not able to provide data to parameterise, force and test models beyond certain levels of detail and complexity. Feeding is one such key processes which regulate mass and energy transfer between organisms and trophic lev els, and ecosystem models cannot deliver reliable predictions unless the data input on these two processes have adequate resolution. In most predictive models today, suspension feeding is often reduced to a simple function of food concentration of homoge neous food items, which often fail to reflect empirical observations, hence the predictions become unreliable. A key step to advance in this field is therefore to establish high resolution and precision feeding data as input in the models, and empirically verify their predictive capability. Here we use the mussel Mytilus edulis as a model species. We propose to link conventional feeding experiments with high-resolution molecular tools to determine mussel feeding- and selection rate of natural seston, incl uding toxic and non toxic phytoplankton, in order to parameterize the variations inherent in the Dynamic Energy Budget (DEB) growth model. The DEB model is a generic model applicable to several marine species, and it has potential application in ecosystem modelling systems. The project therefore also complies with providing significant contributions to the development of parts of ecosystem modelling systems.