Food-Disease-Gene Networks Visualization And Improvement
Oral Defence Date:
Assoc. Prof. Hui Yang & Prof. Ilmi Yoon
The past decade has seen a sharp increase in nutritional genomic publications, making it unrealistic for researchers to analyze manually those papers to arrive at a conclusion. Existing applications focus pri-marily on relationships between proteins and genes for a single task such as named entity recognition or relationship extraction. However, these systems provide limited user interaction and require significant mental effort. We had previously proposed a holistic application to extract verb-based relationships among foods, diseases, chemicals, genes and proteins to build a food-centric network. This thesis describes the maintenance and enhancement in previously implemented visualization interface that allows users to interact visually with the extracted food-centric networks. Also, it illustrates the new features added to the application as well as the refinement of entity recognition and binary relationship extraction approaches. Last, it discusses one version of manual annotation carried on as an evaluation for other modules. The main contributions of this work include a well-functioned visualization application that allows users to explore the gene-food-disease relationship with an improved user interface and new functions like filtering by study parameters.
Visualization, annotation, gene-food-disease network