Lynette Hirschman, The MITRE Corporation Co-authors: Scott Mardis, Cheryl Clark, Kevin Cohen Habitat-Lite (aka EnvO-Lite) is a set of structured... » More

Lynette Hirschman, The MITRE Corporation Co-authors: Scott Mardis, Cheryl Clark, Kevin Cohen Habitat-Lite (aka EnvO-Lite) is a set of structured terms designed to facilitate capture of high-level information about habitat and sample source metadata for genomics and metagenomics samples. It is designed to be light-weight and compact. Habitat-Lite terms are drawn from the full EnvO ontology and are made avail able by the EnvO Consortium in the EnvO-Lite-GSC OBO file. Habitat-Lite terms include high level terms (e.g., terrestrial, marine, freshwater, air, organism-associated) and more specific terms (e.g., soil, sediment, hot spring). These terms were initially selected by Dawn Field (NERC, Oxford), in consultation with domain experts. The terms have been evaluated for coverage and ease of use in capturing relevant information from Genbank isolation source entries, and GOLD HABITAT and ISO LATION entries. We have recently implemented a tool to map automatically from free text phrases into Habitat-Lite terms; we estimate that this tool is capturing 70% of the high level terms correctly, based on comparison to expert manual annotation done by Renzo Kottmann and Pier Buttigieg from the Max Plank Institute, Bremen. Our next steps are to elicit additional use cases from the genomics and metagenomics communities, to develop guidelines for adding terms to Habitat-Lite to capture distinctions critical to these use cases, and to enhance the mapping tool based on these inputs. *This work has been funded under NSF Small Grant for Exploratory Research IIS-0746650.

« Hide