ICBO_2018_67: Vocabularies, Ontologies, APIs, and Formats for Heterogeneous High Throughput Crop Phenotyping Data

TitleICBO_2018_67: Vocabularies, Ontologies, APIs, and Formats for Heterogeneous High Throughput Crop Phenotyping Data
Publication TypeConference Paper
Year of Publication2018
AuthorsLeBauer, D, Willis, C
Conference NameInternational Conference on Biomedical Ontology (ICBO 2018)
Date Published08/06/2018
PublisherInternational Conference on Biological Ontology
KeywordsAPIs, Crop Plants, Environment, High Throughput Crop Phenotyping, High Throughput Phenotyping, Ontology, phenotype

The automation of sensors to measure crops is rapidly increasing in the size and complexity of the data that are available to scientists. The higher resolution, frequency, and diversity of information can enable new scientific discoveries and engineering approaches required to increase the efficiency and sustainability of our agricultural production systems. However, the size and complexity of these data makes it difficult to derive insights and actionable inference. These data often require contextual metadata including plant taxonomy, soil and weather, and agronomic and experimental conditions. While it is easy to agree on the principles of making data FAIR (findable, accessible, interoperable and reusable), it is a more substantial challenge to engineer such data in the absence of clear guidance. The landscape of standards, formats, vocabularies, and ontologies (hereafter, interfaces) that exist is difficult to navigate, and comprehensive specifications that cover the scope of crop phenotyping data are absent. Existing standards and conventions provide only a patchwork of coverage, making it difficult to standardize software and data interfaces. The TERRA Reference phenotyping platform (TERRA REF) is building a suite of open data and software to support advances in the use of crop sensing for breeding and precision agriculture. Although our data are large and diverse, we seek to make it FAIR, useful, and usable to the community of users who represent many science and engineering domains, including computer science, robotics, physics, and biology. This talk will describe our effort to identify, prioritize, and implement interfaces to these data, our efforts to build a community around shared data and software, and our focus on supporting existing software pipelines. It will conclude with a request for feedback on the role of ontologies can play in building a coherent interface to heterogeneous data.