Building the farm to fork digital highway: Integrated food ontology development for agricultural, food science and public health domains
This is a one day workshop on September 16th, 2020 presented as part of the International Conference on Biomedical Ontologies (ICBO) presence within JOWO 2020 in Bolzano, Italy (see https://icbo2020.inf.unibz.it/accepted-workshops/). A call for papers is coming soon. Contact: Damion Dooley, email@example.com
Controlled vocabulary standardization efforts covering agricultural and food domains are evolving since their inception decades ago thanks to the mandates and continued support of institutional caretakers. Popular examples are FoodEx2, the European Food Safety Authority (EFSA) food classification and description system, the Global Language of Business GS1 product categorization scheme, the EUROFIR promoted LanguaL food composition thesaurus, the UN Food and Agriculture Organization’s AGROVOC SKOS-based vocabulary, and its support of the International Network of Food Data Systems (INFOODS) vocabulary for food nutrition testing. These vocabularies are used in a growing interconnected food database landscape but suffer from format issues like textual or spreadsheet formats, unresolvable identifiers, and inconsistent category semantics.
Ontologies are new entrants into the food domain, bringing a wave of Semantic Web technology and philosophy to bear on the issue of data sharing and modelling of food-related activity and research which are becoming critical in the face of rapid change to our environment and anthroposphere. Examples range from BBC’s Food Ontology, driving its culinary media universe, to recently published research laboratory initiated ontologies like OBOFoundry members FoodOn, the Food Biomarker Ontology (FOBI), and the Ontology for Nutritional Studies (ONS), and related ontologies like the Medical Action Ontology (MAxO) and the Environmental Conditions, Treatments and Exposures ontology (ECTO) that are under development. Underpinning these mid-level, model-focused ontologies are environmental, chemical, biological, anatomical, disease and phenotype ontologies.
Academic, agricultural and public health agencies are considering the benefits and complexities of adopting ontology in their research and data management and reporting infrastructure. How can ontologies interface to legacy datasets and online databases described by existing vocabularies? What vocabulary, tool ecosystem and data models are needed to correlate agricultural treatments, nutritional data, eating patterns, biomarkers, pathogens, and phytochemical levels with disease and health phenotypes? This workshop seeks to define the coverage of the different ecological, agricultural, nutritional, dietary, public health, one health surveillance, food security, and trade domains that food-related ontologies are modelling, and the use of data translation tools for bringing legacy data into the ontology fold.
Workshop will have 2 hour sessions split into presentations of relevant ontologies / projects and their scope in the first hour, in 10-15 minute slots + 5 for a few questions, followed by 1 hour discussion and a break as schedule allows.
Session: In the food: Food component modelling
Modelling biochemical food components, and how they are taken in the body through dietary eating patterns. Food biochemistry of plants and animal products, including their biotransformation from fungi, and bacteria. Modelling of bioactive food related compounds including nutrients, additives, and contaminants, as well as diet terms and regimens. Food allergen and other linkages between food components and disease phenotypes. Finally modeling the role of microbial fermentations from pure culture or microbial communities and the metabolic interactions between the components of the microbial web will be a task.
Session: In the body: Modelling metabolic transformation of food and exposure to food-borne toxicants
Can we model the complex web of metabolic transformations leading from food components to beneficial bioactives for health maintenance or improvement? This session welcomes contributions dealing with ontology modeling of bioactive molecules derived and/or modified from food components by human metabolism, by gut microbiome metabolism, and their synergy. We include xenobiotic metabolism and exposure through food as well.
Session: In the field: Agricultural production modelling
How should agricultural field studies be modelled, and can this adapt to commercial farm data collection? One of the challenges with agricultural field modelling is having to rely on estimated (or subjective) vs actual data since proper measurements occur in limited circumstances. A second challenge is collecting data from farms with heterogeneous practices in large, remote areas in a way respecting privacy. Use of agricultural chemicals and their potential inclusion in food products must also be considered. Finally, vocabularies for product vary by region, language and industry making vocabularies a necessity for research, food traceability and correct product labelling.
Session: In the agency: Ontology ecosystem to support research and operational adoption
The promise that ontology will fix data interoperability carries possible failure points along a two-directional dependency chain that has term curation at one end, then dataset curation by trained agency staff according to selected vocabulary and design patterns, then the provision of that data to federated querying or consolidated databases, and lastly, users tasked with querying and learning about what are potentially very complex data structures. Are success stories accumulating as ontology identifiers surface within a diversity of databases?