Making sense of fuzzy data: new approach promises clear insights for tackling disease outbreaks in farmed fish


October 30, 2019 Facebook Twitter LinkedIn Google+ Aquaculture News,News-Global



A quick, cost-effective, Cloud-based solution to one of the biggest challenges to fish farming could soon be accessible to farmers, scientists and policy makers if a new project gets the go-ahead.

Aquaculture is one of the fastest-growing food production sectors in the world and will be essential to meeting the 40% increase in demand for food fish expected by 2030, according to the Food and Agriculture Organization of the United Nations. The sector is already providing incomes and new jobs for women and men of all ages particularly in Africa, through fish production, processing and trade, and helping diversify diets that often lack in essential micronutrients.

But a key challenge to aquaculture is outbreaks of disease. Tilapia lake virus and streptococcal infections, among others, can quickly wipe out farmed fish and are becoming more common as a result of overstocking, poor hygiene practices, and insufficient controls at hatcheries and farms. Managing outbreaks is often hindered by limited knowledge of the pathogens, which in turn contributes to anti-microbial resistance, as farmers use the wrong kinds or quantities of antibiotics on their fish.

Now, a new approach to disease diagnostics could change this. By bringing together years of fish pathogen data and combining it with new pathogen DNA sequences from infected fish, it aims to build a Cloud-based service that accurately diagnoses a range of fish diseases. Management advice will then be sent back to farmers, extension workers, hatcheries and quarantine officers in near-real time, and will also provide scientists with information to help them develop vaccines. If successful, the approach could be applied to a range of pathogens affecting all farmed animals.

The team behind the concept is among nine finalists in the Inspire Challenge 2019, organised by the CGIAR Platform on Big Data in Agriculture. The competitive award funds innovative proposals that can help bring the power of big data to small-scale food producers. Each finalist will pitch their idea to a panel of experts at the Big Data Convention in Hyderabad, India this week, with the five winners each receiving a grant of US$100,000 to develop and test their concept over the following year.

“Rapid genome sequencing provides all of the information we need to make informed decisions on disease control without the need for pathogen-specific tests and tools that might be very expensive,” said Dr. Andrew Barnes of The University of Queensland (UQ), who is part of the team behind the WorldFish-led proposal. “The sequences enable you to infer the origin of a pathogen; see how it is evolving and moving through different environments and across international borders. Most importantly, it allows you to identify possible antimicrobial resistance genes and factors relevant to vaccine formulation. With this kind of knowledge we can provide very specific advice on how to control and prevent outbreaks.

Until recently, however, sequencing hardware has been expensive and available only in  specialist labs. This has changed in the last few years with cheap, user-friendly, pocket-sized instruments enabling sequencing to be carried out locally and quickly. The challenge is that these tools generate ‘fuzzy’ data that is difficult to interpret precisely – which is where big data analytics comes in.

The team, led by Dr. Jerome Delamare-Deboutteville of WorldFish, will generate genomic sequences from pathogens of tilapia and carp, the world’s most important farmed fish. These will be sourced from a range of partners, including those in Bangladesh, Malaysia and Thailand, and used to combined with datasets held at UQ and on public databases from around the world in one common online platform. Next, software and machine learning tools developed by project partner Wilderlab – a New Zealand-based analytics firm – will extract specific pathogen information from the datasets. This will enable scientists to generate precise management recommendations that can be applied straight away.

Dr. Delamare-Deboutteville of WorldFish said: “Rapid diagnosis combined with the effective management can mean the difference between a fish farmer losing all her stock within days of a disease outbreak and consistently producing a bumper harvest. Multiply that across one of the world’s fastest-growing food production sectors, and all of the vital nutrients, jobs and opportunities it provides and you could have an enormous impact on well-being in developing countries.

“We’re in the midst of a genomics revolution, and our approach combines the formidable power of genomics with the huge potential of big data analytics to get the best of both worlds. If successful, it could represent a quantum leap in responding to the major production constraints for farmed fish, and could be applied to rapid disease detection, response and prevention in all farmed animals around the world.” 

“The key to all of this is good, reliable data, powerful processing, expert partners, and a belief in the power of disruptive thinking that’s at the core of the Inspire Challenge. Like so many game-changing ideas, ours is ultimately quite simple, but it could close a major production gap for many small-scale producers of fish and other livestock.” 

Click to read the full project proposal, Rapid genomic detection of aquaculture pathogens.