Inference of Functions, Roles, and Applications of Chemicals Using Linked Open Data and Ontologies

• Published in: Semantic Technology Vol. 11341; pp. 385 - 397
• Main Authors: Kushida, Tatsuya, Kozaki, Kouji, Kawamura, Takahiro, Tateisi, Yuka, Yamamoto, Yasunori, Takagi, Toshihisa
• Format: Book Chapter
• Language: English
• Published: Switzerland Springer International Publishing AG 2018; Springer International Publishing
• Series: Lecture Notes in Computer Science
• Subjects: Chemical; Disease; Drug; Inference; Knowledge graph; LOD; Ontology; RDF; SPARQL
• ISBN: 3030042839; 9783030042837
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-030-04284-4_26

A simple method to efficiently collect reliable chemical information was studied for developing an ontological foundation. Even ChEBI, a major chemical ontology, which consists of approximately 90,000 chemicals and information about 1,000 biological and chemical roles, and applications, lacks information regarding the roles of most of the chemicals. NikkajiRDF, linked open data which provide information of approximately 3.5 million chemicals and 694 application examples, is also being developed. NikkajiRDF was integrated with Interlinking Ontology for Biological Concepts (IOBC), which includes 80,000 concepts, including information on a number of diseases and drugs. As a result, it was possible to infer new information on at least one of the 432 biological and chemical functions, applications and involvements with biological phenomena, including diseases to 5,038 chemicals using IOBC’s ontological structure. Furthermore, seven chemicals and drugs, which would be involved in 16 diseases, were discovered using knowledge graphs that were developed from IOBC.