Thumbnails:
List:
Year:
Category:
Session:
Poster:
Getting poster data...
William J.R. Longabaugh (Institute for Systems Biology, Seattle, WA USA)
By representing nodes as horizontal lines instead of as essentially zero-dimensional points, BioFabric can sidestep the "hairball trap"; it is able to visualize large networks in an unambiguous, organized, and scalable fashion. The additional degree of freedom provided by "nodes as lines" allows edges, which appear as parallel vertical line segments, to be carefully arranged in the horizontal dimension to maximize clarity. In particular, the "edge wedges" created by this technique provide a visually distinctive, and directly comparable, signature of each node's connectivity pattern. This visualization of the Center for Cancer Systems Biology HI2_2011 protein-protein interaction dataset contains 2193 nodes and 3881 edges, and is laid out using BioFabric's default breadth-first layout algorithm. The addition of duplicate "shadow links" above the main diagonal doubles the edge count, but provides each node with its full complement of incident edges in its dedicated labeled node zone. The full network is shown at lower left, while the main image is a detailed view of the upper left corner of the network.