Partitioned Layout of Protein Interaction Networks

초록

Existing visualization studies in protein interactions suggest that the nature of protein interaction data requires a new graph layout method for protein interaction networks. Protein interaction data can be characterized as follows: - The data yields a nonplanar graph with a large number of edge crossings that cannot be removed in a two-dimensional drawing. - Proteins have a very wide range of interacting proteins within the same set of data, so a graph visualizing the data contains nodes of very high degree as well as those of low degree. - When visualized as a graph, the data yields a disconnected graph with many connected components. - The data often contains protein interactions corresponding to self-loops. Considering these characteristics of protein interaction data, we have developed a new force-directed layout algorithm for drawing protein interactions in three-dimensional space. Our algorithm partitions nodes into three groups based on their interacting properties: a biconnected subgraph in the center of a graph ($V_3$), terminal nodes at the outermost region ($V_1$), and the rest in between them ($V_2$). These partitions are identified from the outer partition ($V_1$) to inner one ($V_3$) and are placed in reverse order. Experimental results show that our algorithm efficiently generates a clear and aesthetically pleasing drawing of large-scale protein interaction networks and that it is much faster than other force-directed layouts.