Week 3: Protégé 5 and OWL

Moving into the third week of the project I pulled back a bit and did some background reading. The issue with the Protégé 5 ontology imports I decided was an issue stemming from my ignorance with regards to the software and terminology therein. To remedy this I read the book Semantic Web for the Working Ontologist: Effective Modeling in RDFS and OWL 2nd Edition by Dean Allemang and Jim Hendler. This book provided me with a detailed explanation of the Semantic Web. The text is filled with examples and by the end I felt fluent in the terminology. Following reading the book by Allemang and Hendler I returned to troubleshooting the Protégé 5 ECSO (Ecosystem Ontology) import. Previously, the ECSO import to Protégé 5 had been unreadable with only the ontology ID displayed (ex. TAXA_000002860, ECSO_00002629, etc.) not the class rdfs label (ex. mollusc, methane, etc.). I was able to resolve this issue by going into the Preferences in Protégé and under the New Entities tab selecting “end with auto-generated ID for entity label”, then specifying that it should be the “same as label render” and changing the language to “en” for English. Following this, I changed the View in Protégé to “Render by annotation properties” and then “Render by label (rdfs:label).” Of the 1,959 classes within the ECSO all but 52 classes now display their proper rdfs label. The 52 that did not properly render lack any annotation information and are “naked terms.” While there are still issues with these 52 imports there does not appear to be a pattern in their OWL model construction. Currently the unlabeled classes consist of 3 ECSO, 16 ENVO (Environment Ontology), 32 PATO (Phenotypic Quality Ontology) and one UO (Units of Measure Ontology) class. I will have more on this mystery next week!

Another goal this week was to evaluate whether the the queried search terms regarding carbon cycling on the Arctic Data Center are present in the ECSO ontology. The query term list includes those from Chapin et al., 2006 (mentioned last week), new terms have been added from results found within the database during query searches, and select terms from the Glossary of Permafrost and Related Ground-Ice Terms  from the National Research Council of Canada and the International Permafrost Association. Of the 51 queried terms there were a total of 2486 search results within the ADC database. Within ECSO there are currently 2 of these 51 queried carbon-cycling related terms within the ontology. The two terms are “carbon” and “methane” which account for 21% of the total search results and are clearly essential terms for the ADC carbon-cycling ontology. Moving forward next week I will be focusing on knowledge modeling itself of carbon-cycling terms within ECSO in Protégé 5.

Leave a Reply

Your email address will not be published. Required fields are marked *