BioNET : traversing and visualising biological reaction networks.

Abstract

Accurately traversing biological reaction networks, or metabolic pathways, is a multifaceted problem. It is affected by issues involved in representing the pathways, textually as well as graphically. Furthermore the traversal is in itself a combinatorial optimisation problem. To this day not much work has been done in trying to solve this issue. This thesis targets the traversal problem through the use of the Ant Colony Optimisation metaheuristic. A three dimensional layout of the graph is rendered so that results can be animated. Furthermore a heuristic based on shadow casting principles is presented, providing a means by which to observe the animation in an optimal manner. This will effectively target the problems involved in visualising such networks. It will be shown how the traversal algorithm proves to be an effective approach to this combinatorial optimisation problem. Furthermore, the simple concept of shadow casting is shown to achieve optimal performance in visualising the network.