Use of ontologies to document characters

RAMÍREZ, M. J.; MADDISON, W.; CODDINGTON, J. A.

Tucumán, Argentina

Congreso; . XXVII meeting of the Willi Hennig Society; 2008

Instituto Miguel Lillo

As scholarship advances, no single person can know all relevant anatomical structures, much less all the interpretations of those structures.  A person scoring cells in a dataset needs at least a succinct documentation for each character state, especially in a multi-author project.  Composing such character documentation is however not trivial.  In large-scale, encyclopedic projects, the list of characters may easily scale up to hundreds or thousands, hence the associated documentation becomes monograph-sized by itself.  There are several factors by which proper documentation of characters is rarely done, among others: (1) it is hard for phylogeneticists to keep pace with publications relevant to characters, especially on functional morphology and experimental development; (2) character documentation usually duplicates previous work in a large extent; (3) it is hard to transform the monumental work of character documentation into a regular publication; (4) if the documentation is published, it is hard to convert and maintain free text publications (paper-based or digital) into machine readable formats to document future phylogenetic datasets.  It turns out that there is a simple way to convey community-based repositories of biological data into phylogenetic datasets, thus alleviating a significant part of their documentation.  Dataset editors such as Mesquite and MX can make use of ontologies to document characters.  We present an application of this idea in Mesquite, using ontologies in Obofoundry.  These ontologies are structured controlled vocabularies with definitions, synonyms, and relationships between terms, curated by a community of experts, machine readable, and exposed in public repositories.  The basic idea is to associate characters with the relevant terms in an ontology, thus allowing Mesquite to display the information in the ontology relevant to each character.  By this simple operation, a large part of character documentation and related functionality can rely on an external repository, opening channels for interaction between communities working with model organisms and with diversity.