Annotation of Crop Data with Ontologies to Support Fitness for Use

Modes of plant trait data collection are ranging from the field data capture, to greenhouse and high throughput phenotyping platforms. All modes associate the trait with the environment and the crop management technics to support spatial and temporal modelling of trait and crop performance. The fitness for use of plant trait data includes the harmonization and quality of their annotations from capture to publishing. The annotations must enable proper interpretation of data, an objective that the CGIAR Crop Ontology and Agronomy Ontology attempt to support with the creation of ‘standard variables’ composed by a ‘a trait or a property + a method + a unit or a scale’. Standard variables support the creation by breeders or agronomists of electronic fieldbooks that will directly annotate observations, without requiring extra effort from the data collector, and secure the transfer into databases and repositories. These annotations will make data available through the new standard Breeding Application Programming Interface (BRAPI; http://docs.brapi.apiary.io/#).

To expand their data search and discovery, scientists also need over-arching ontologies where concepts are independent of the species and this is addressed by the set of reference plant ontologies of the Planteome (www.planteome.org).

We will explain how projects for data management, involving CGIAR, approach the use of a common pool of standard variables, ontologies, and controlled vocabularies like the Breeding Management System of Integrated Breeding Platform, the Next Generation Breeding Databases of the CGIAR Root, Tubers and Banana Research Programme, the Global Agricultural Trial Repository (www.agtrials.org), the Agricultural Model Intercomparison and Improvement Project (AgMIP), and the CGIAR Big Data Plaform project.