Degree Type

Dissertation

Date of Award

2006

Degree Name

Doctor of Philosophy

Department

Electrical and Computer Engineering

First Advisor

Julie A. Dickerson

Abstract

A combination of graph layouts in 3D space, interactive computer graphics, and virtual reality (VR) can increase the size of understandable networks for metabolic network visualization. Two models, the directed graph and the compound graph, were used to represent a metabolic network. The directed graph, or nonhierarchical visualization, considers the adjacency relationships. For the nonhierarchical visualization, the weighted GEM-3D layout was adopted to emphasize the reactions among metabolite nodes. The compound graph, or hierarchical visualization, explicitly takes the hierarchical relationships like the pathway-molecule hierarchy or the compartment-molecule hierarchy into consideration to improve the performance and perception. An algorithm was designed, which combines the hierarchical force model with the simulated annealing method, to efficiently generate an effective layout for the compound graph. A detail-on-demand method improved the rendering performance and perception of the hierarchical visualization. The directed graph was also used to represent a sub-network composed of reactions of interest (ROIs), which reveal reactions involving a specific node. The fan layout was proposed for ROIs focusing on a metabolite node. The radial layout was adopted for ROIs focusing on a gene node. Graphics scenes were constructed for the network. The shapes and material properties of geometric objects, such as colors, transparencies, and textures, can encode biological properties, such as node names, reaction edge types, etc. Graphics animations like color morph, shape morph, and edge vibration were used to superimpose gene expression profiling data to the network. Interactions for an effective visualization were defined and implemented using VR interfaces. A pilot usability study and some qualitative comparisons were conducted to show potential advantages of stereoscopic VR for metabolic network visualization.

DOI

https://doi.org/10.31274/rtd-180813-16514

Publisher

Digital Repository @ Iowa State University, http://lib.dr.iastate.edu/

Copyright Owner

Yuting Yang

Language

en

Proquest ID

AAI3229141

File Format

application/pdf

File Size

107 pages

Share

COinS