A pragmatic path forward for integrating phenotype and trait data using ontologies

Phenotypes are hard to deal with. Unlike genotypes, which can be described using sequences of characters, phenotypic data are conditional and non-standardized and therefore normally described using text. Even quantitative phenotype data (e.g., plant height in cm) require additional explanation such as how and when they measured to be understood fully. With the growth of new automated phenotyping and trait measurement methods, we are in desperate need of scalable, standardizable methods for annotating and storing phenotype and trait data, if we ever want that data to be discoverable, computable, and re-usable. Phenotype and trait ontologies (well-defined sets of terms with logical relationships linking them) hold the promise of standardizing phenotype data, but only a very small percentage of researchers use ontology terms to describe their data. In this talk, I will describe some recent efforts that make use of ontologies for describing phenotypes and lay out a roadmap for how to build on this work for more scalable data systems. Topics will include cross-species comparisons of mutant phenotypes, managing biodiversity trait data, and linking ontologies to metadata.