Scientists use artificial intelligence to construct food webs
30 December 2011
Scientists at Rothamsted Research, which receives strategic funding from BBSRC, in collaboration with Syngenta, Imperial College London and the French National Institute for Agricultural Research (INRA) have developed an advanced technique, using a branch of artificial intelligence called Machine Learning that could greatly improve our understanding of the interaction between food and the environment.
The paper, published today in PLoS One, explains how bioinformatics can be used to construct food webs from ecological data, for example allowing us to better understand how agricultural management affects the agricultural ecosystem. Food Webs describe 'who eats whom' within the ecosystem, by linking together species that feed on one another into multiple food chains that explain how food energy moves through the ecosystem.
Dr David Bohan, the lead author said that "this approach has great power and could be used to extend and test our current theories of ecosystem dynamics and function." He added "we believe it could be used to support the development of a predictive theory of ecosystem responses to environmental change."
This will help us explain how we affect biodiversity and the benefits of natural ecosystems when we manage our crops. But this technique is not limited to agricultural ecosystems as it can be used to predict the environmental functions, or 'jobs' that could be lost as the demands on biodiversity are affected by the "perfect storm" of climate change, fuel and water scarcity and population increase.
The technique therefore could be of significant importance and we try to find solutions to food, water and energy security in the face of climate change.
With the current world population of 7 billion projected to reach 9.3 billion by 2050, an average increase of over 160,000 people every day, it has never been more important to understand food webs in the ecosystem as we drive forward to develop sustainable agricultural systems to increase our food production with a reduced environmental impact.
Machine Learning allows computers to learn food webs based on existing data. The technique described by Dr Bohan and his colleagues uses a logic-based approach called A/ILP which has generated plausible and testable food webs from a national-scale sample of invertebrates from arable fields in Great Britain.
Notes to editors
The article entitled "Automated Discovery of Food Webs from Ecological Data using Logic-based Machine Learning" was published in PLoS ONE on 12/20/2011 and is available online at http://dx.plos.org/10.1371/journal.pone.0029028
About Rothamsted Research
Rothamsted Research, which receives strategic funding from the Biotechnology and Biological Sciences Research Council (BBSRC), is the longest running agricultural research station in the world, providing cutting-edge science and innovation for around 170 years. Its mission is to deliver the knowledge and new practices to increase crop productivity and quality and to develop environmentally sustainable solutions for food and energy production. More information can be found at www.rothamsted.ac.uk
BBSRC invests in world-class bioscience research and training on behalf of the UK public. Our aim is to further scientific knowledge, to promote economic growth, wealth and job creation and to improve quality of life in the UK and beyond.
Funded by Government, and with an annual budget of around £445M, we support research and training in universities and strategically funded institutes. BBSRC research and the people we fund are helping society to meet major challenges, including food security, green energy and healthier, longer lives. Our investments underpin important UK economic sectors, such as farming, food, industrial biotechnology and pharmaceuticals.