Stable isotopes versus DNA-based tools: can we use these methods to identify commercial fish stocks?

The Seafish Common Language Group meeting

Wednesday June 27, 2018, I had the chance to take part in the Seafish Common Language Group (CLG) meeting in London where fascinating topics were discussed ranging from the role of science-based methods in product integrity to addressing the risks and mitigation measures of marine litter.

I would need more than one blog post to talk about the presentations and engaging discussions that followed. Here, I wanted to address a topic that is closely related to what I will be focusing on during my PhD thesis, the traceability of fish stocks using science-based tools. I was therefore particularly curious to hear what Stewart Whitehead from Oritain and Clive Truman from the University of Southampton had to share as I knew that they would be talking about the use of geochemical methods for the traceability of fish stocks.

What’s seafood mislabelling anyway?

You’d have to have your head buried in sand not to have heard of the issue of seafood mislabelling that has been extensively discussed both in the scientific literature and the media over the last decade. Ongoing advances in the field of genetics have enabled the use of DNA-based tools for forensic science and have allowed scientists to identify fish species that come under the rather perplexing form of a filet, a sushi, a can, or a fresh, dried, frozen, or cooked steak. Thanks to these methods, it has been possible to report and denounce the extensive amount of mislabelling and species swapping that occurred around the world both in the retail sector and in restaurants. Despite the plethora of articles on that matter, it has been difficult to identify where in the supply chain does this “swapping” occurs but one thing is for certain, it is happening, whether voluntarily or not.

Why traceability of fish stocks?

Until now, most if not all studies have focused on species mislabelling, but in truth, mislabelling is not unique to the species entity. Stock mislabelling (i.e. mislabelling the provenance of a specimen) is something that is currently much more difficult to demonstrate using available technological tools and yet, the motives to pretend that a fish comes from a stock that is sustainably managed as opposed to a stock that is depleted are likely prevalent in the seafood industry.

How reliable are stable isotopes and DNA-based tools for the identification of fish stocks?

While Stewart’s presentation was quite optimistic regarding the use of stable isotopes in the industry and included many successful examples, Clive’s offered a more nuanced picture highlighting the remaining uncertainties and the many difficulties to overcome if one is to use stable isotopes for seafood forensics. The stable isotopes of Carbon and Nitrogen used in marine ecosystems do not exactly give information about the geographic location of an organism but rather about its position in a complex food web, which can then be used to draw spatial gradients and estimate its provenance. Bearing that in mind, it can be a relatively simple or an extremely complex task to identify where an individual fish comes from, and it will largely depend on the way the question is formulated. If the question is “did this fish come from this specific place?”, the fact that we have a reference location offers methodological advantages. If the question however is “where did this fish come from?”, the fact that we need to consider all possible location sources severely complicates the task.

As an example, Clive mentioned a pilot study on cod from the North East Atlantic. Here, there was a defined number of four specific locations and Atlantic cod (Gadus morhua) were sampled from these locations. For some locations such as the Barents sea, 80% of the cod could be assigned back to their population of origin. In other locations such as Iceland waters which are transitional waters, the analysis showed very poor isotopic characterisation, making it impossible to trace the individuals back to that specific location. Considering the relatively good precision of this study, the following question is, can this be applied in the seafood industry? Clive presented a study that did just this and that successfully used these methods to look at where marketed fish came from (such as fish & ships cod in the U.K. which unsurprisingly appeared to have come from the Barents sea).

This pilot study invariably made me think of a similar study using DNA-based tools that I was to later present at the meeting. In a publication from 2012, Neilsen and colleagues were able to attribute 98-100% individual cod back to their population of origin. Another study published in 2018 by Drinan and colleagues assigned 88% of individual Pacific cod (Gadus macrocephalus) back to the sample from which they were captured (for some samples, 100% of the individuals were accurately assigned back to the catch sample).

As I mentioned earlier however, a question can be much more complex to answer if the possible places from which a product may have come from are unknown. In that case, researchers need to make predictive spatial models and will end up with a map showing how likely it is that each individual fits within given geographical cells. A stable isotopes pilot study on scallops showed that this method functioned well for that specific sedentary organism. Clive highlighted however that this stable isotopes-based spatial modelling method cannot currently be applied to a global scale because we do not have a good enough database to build a global model.

Aside from the issue of scale it should also be recognized that some animals are better suited to this method because of the inherent way they assimilate stable isotopes. Animals that stay in a given location over the course of their lives are ideal candidates. On the other hand, stable isotopes analysis will not be suited to higher trophic level and highly migratory oceanic fishes such as tuna and this can be true both for stable isotopes and DNA-based methods. The success of DNA-based methods for stock identification depends on a population’s degree of reproductive isolation (i.e. A highly mobile species who breeds with individuals from across the world will not display population-specific polymorphism). On the other hand, if a species is highly mobile but relatively isolated when it reproduces (i.e. Always comes back to the same location to reproduce), DNA-based tools will successfully differentiate between populations whereas stable isotopes analysis would not be able to.

Science-based methods for stock identification will be equally hindered by the fact that stock boundaries are not always synonymous with population boundaries. Populations themselves can prove flexible in terms of geographical preference, expending or shrinking their boundaries as a response to external environmental and ecological variables, overlapping with other populations, and regularly changing their patterns of migration and reproduction. Therefore, as Clive put it, we need a good understanding of the animal’s ecology and its interaction with the environment before we can say whether any of these methods are sensible to use in order to trace individuals back to a stock of origin. During the next few months, Clive Turman, Kathy St. John Glue, Stefano Mariani, Ilaria Coscia, and myself will attempt to elucidate whether these methods are reliable enough to be used as a tool for the implementation of regulation in the seafood industry or whether they can currently solely serve as a deterrent.