Tatsuya Kushida

Event ontology: a pathway-centric ontology for biological processes.

Event ontology is a new biomedical ontology developed to annotate pathway components in a pathway database. It organizes the concepts and terms of sub-pathways, pathways, biological phenomena, experimental conditions, medications, and external stimuli appearing in biological pathways (e.g. signal transduction, disease-, metabolic-, molecular interaction-, genetic interaction pathways, etc.). Concepts in the Event ontology are extracted manually from scientific literature. Each term has links to external databases such as Gene Ontology, Reactome, KEGG, BioCyc, and PubMed.

Collection of Disease Networks by Hybrid Curation Method and the Application for Pathway Analysis

The network information on biological phenomena, such as incurable diseases, which were seldom known until now, was collected from more than 18,000,000 biomedical articles by a hybrid curation method which consisted of a machine curation by an information extraction system, GENPAC, and a manual curation by biologists. So far, 64 kinds of disease networks which comprise nodes of genes, proteins, chemicals, and biological phenomena, and edges of the interaction types, such as “Activate” and “Bind” have been curated by the method. The information extraction of the networks shows high recall and precision rate, and furthermore, the disease networks information can be collected by the hybrid curation more efficiently than by only expert manual curation. It can be considered that the interaction and network information which is collected by the hybrid curation will be utilized for the functional annotation of gene clusters and discovering new subpathways on biological pathway maps.

A preliminary approach to creating an overview of lactoferrin multi-functionality utilizing a text mining method

Lactoferrin is a multi-functional metal-binding glycoprotein that exhibits many biological functions of interest to many researchers from the fields of clinical medicine, dentistry, pharmacology, veterinary medicine, nutrition and milk science. To date, a number of academic reports concerning the biological activities of lactoferrin have been published and are easily accessible through public data repositories. However, as the literature is expanding daily, this presents challenges in understanding the larger picture of lactoferrin function and mechanisms. In order to overcome the “analysis paralysis” associated with lactoferrin information, we attempted to apply a text mining method to the accumulated lactoferrin literature. To this end, we used the information extraction system GENPAC (provided by Nalapro Technologies Inc., Tokyo). This information extraction system uses natural language processing and text mining technology. This system analyzes the sentences and titles from abstracts stored in the PubMed database, and can automatically extract binary relations that consist of interactions between genes/proteins, chemicals and diseases/functions. We expect that such information visualization analysis will be useful in determining novel relationships among a multitude of lactoferrin functions and mechanisms. We have demonstrated the utilization of this method to find pathways of lactoferrin participation in neovascularization, Helicobacter pylori attack on gastric mucosa, atopic dermatitis and lipid metabolism.

Expanding Science and Technology Thesauri from Bibliographic Datasets Using Word Embedding

The use of thesauri and taxonomies for science and technology information in scientometrics has been attracting attention. However, manual construction and maintenance of thesauri is expensive and requires significant time, thus, methods for semi-automatic construction and maintenance are being actively studied. We propose a method to expand an existing thesaurus using the abstracts of articles from state-of-the-art technological domains with limited structured information. Specifically, we consider a method for properly allocating new terms to the hierarchical structures of an existing thesaurus using rapidly evolving word embedding. In an experiment, word vectors of 500 degrees are constructed from 567,000 biomedical articles and are clustered after dimension reduction using principal component analysis. Then, semantic relations are estimated based on the spatial relations between the new term and any of the terms in the thesaurus. We then conducted a comparison of the results obtained from three experts. In future, we will develop a recommendation system for new terms related to the existing terms to support semi-automatic thesaurus maintenance.

Refined JST Thesaurus Extended with Data from Other Open Life Science Data Sources

We are developing a refined Japan Science and Technology (JST) thesaurus with thirty-five relations to enable description of rigorous relationships among concepts. In this study, we prepared an environment for performing SPARQL queries and evaluated the JST thesaurus in the life sciences by comparing query results with the originals. Based on the results of the investigation, we constructed a fibrinolysis network from the thesaurus as a collection of concepts connected with fibrinolysis within three steps, and we discovered that fibrinolysis was associated with fifty-four concepts, including sixteen diseases and twelve physiological phenomena. Subsequently, using the sub-classified relations, we divided the sixteen diseases into two diseases that developed after fibrinolysis progressed, seven diseases that shared common molecules in the development mechanism with fibrinolysis, and other associated conditions. Furthermore, we mapped concepts between the JST thesaurus, ChEBI, and Gene Ontology by matching the labels and synonyms. As a result, we could integrate the fibrinolysis network with thirty-seven chemicals, including four antifibrinolytic agents and twenty-seven human gene products that can regulate fibrinolysis. Thus, we were able to handle the information relating to a series of molecules, molecular-level biological phenomena, and diseases by integrating the refined JST thesaurus with information regarding chemicals and gene products from other resources.