Ontology-based metabolic pathway prediction using saccharomyces cerevisiae data from genbank, ecocyc and kegg

Abstract

Nowadays, metabolic pathway prediction is the focus of numerous drug discovery researches and is central to the stage of many biopharmaceutical and genomic companies. The identification and validation of drug targets depends critically on knowledge of the metabolic pathways in which potential target molecules operate within cells. To understand the cellular function, most scientists and biologists study parts of metabolic pathways that contain various types of pathways and large volume of knowledge including genes, enzymes, chemical compounds, and reactions that interlinked with each other. Thus, in order to provide better access to relevant knowledge, the knowledge of metabolic pathway should be conceptualized and formalized using appropriate knowledge representation technique. Currently, there is no ontology which is developed specify in metabolic pathway domain. Therefore, the main objective of this research is to develop an ontology-based representation for metabolic pathway to represent and to describe the concepts in the metabolic pathway domain and the relationships among them. Then, the metabolic pathway ontology is manipulated to predict and analyze metabolic pathways for a target organism using metabolic pathway prediction algorithm. To enhance the efficiency in predicting metabolic pathways, Problem Solving Method approach is proposed to perform the ontology inference by providing the reasoning component to solve the prediction problem of metabolic pathways. This proposed approach is implemented and tested using real data of Saccharomyces cerevisiae from GenBank and pathway reference databases from EcoCyc/MetaCyc and KEGG. This research presents the integration of ontology and metabolic pathway prediction algorithm as a possible solution for predicting metabolic pathways. This approach capable to predict metabolic pathway of S.cerevisiae with 87 percent accuracy compared to 80 percent accuracy using PathoLogic algorithm