Study On Visualization Methods Of Complex Biological Networks

Biological networks play an essential role in systems biology. Rapidly growing network data and versatile research activities promote greatly the development of visualization technique and call for fast algorithms to draw networks and convenient visualization tools to aid intuitively perceiving abstract structures of networks and gaining insights into the functional implications of networks.It is necessary to design suitable algorithms to draw biological networks.General drawing methods such as hierarchical layout, circular layout, force-directed layout are suitable for biological networks of fewer nodes and simple topological structure and usually not adequate to produce satisfactory drawings for complex biochemical networks of large amount of nodes and connections. Grid layout methods draw many concerns for having advantages in generating compact layouts with biologically comprehensible modules of biological networks.The main issues of grid layout methods are the high computational cost and large computer memory requirement, which seriously limit the applications in interactive real-time visualization analyses of complex biological networks. Some improved algorithms proposed by Kato and Kojima tried to speed up layout processes by the method of adding cross costs and biological attributes in cost functions, but they still need relatively long layout time. It is necessary to develop new fast grid layout methods to achieve near real-time visualization for complex biological networks.The visualization software systems are used to display layout results intuitively and operate interactively through the graphical user interfaces. The systems such as Cytoscape,VisAnt, Cell Illustrator, and CADLIVE are the most popular visualization systems recently. Generally speaking, it is very necessary to analyze the data of complex networks during visualization processes for the purpose of integrating and analyzing networks efficiently. However, the mentioned systems are designed for modeling and specific purposes, network analyses related numerical utilities are not provided, therefore, it is not very convenient for researchers to analyze the data of complex biological networks.In the graphical user interfaces of complex biological networks, large amount of nodes and uneven connections usually cause highly visual complexity. How to display node annotation information suitably in limited visual space remains an essential problem. A good soultion to the problem will help users learn the information and characteristics of networks completely and intuitively.This paper presents the following works based on the above-mentioned analyses:a) A fast grid layout algorithm has been designed and with it, the high quaility layouts can be produced in less computer memory and lower computational cost. The module structures in result drawings are biological understandable. All these advantages enable visualization of complex biological networks in near real-time in interactive visual environment.b) A visualization software system for complex biological networks LucidDraw has been developed. The fast grid layout is implemented as the computation core of LucidDraw and, MATLAB,a convenient, general-purpose network analysis tool is integrated, providing users a convenient way to visually analyze complex networks. An interactive graphical user interfaces for visualization is designed in LucidDraw, in addition, graphical user interface for controlling algorithm parameters is provided. Users can modify layout styles and algorithm parameters as wish. LucidDraw also enables easy incorporation of extra biological information, if available, to influence the output layouts with predefined node grouping features.c) A solution of effective displaying annotation information in the context of drawings of complex networks is presented. Three kinds of labels, i.e., engraved, floating, and mandatory labels, are used to aid users to get node information conveniently with minimum perturbation of the whole drawing. Displaying of engraved labels is dependent on the zoom level, which can draw more labels as possible without causing too much visual complexity.A case study for analyzing P. aeruginosa PAO1 is given to illustrate the application of LucidDraw. With the help of LucidDraw, the relation between virulence factor synthesis and other functional subsystems can be viewed intuitively and the drawings can also provide clues for further investigations to clear the uncertainty for some predicted reactions with unknown assignment of functional subsystems.