Visualizing a Correlative Multi-level Graph of Biology Entity Interactions

In this paper we present a new visualization paradigm to represent and assist the understanding of a correlative multi-level graph, a group of interconnected networks. Such a graph is formed via term association mining, and the visualization paradigm consists of three components: terrain surface visualization units, terrain surface arrangement, and terrain surface correlation. We apply this paradigm to visualize and explore a pair of correlative core cancer terms network and core gene terms network. The results show that our visualization paradigm design is consistent with the derived associations, and is effective in preserving major features as the landmarks in the terrain surfaces.