Biofabric is a visualization technique used to represent networks, and is unique in that it represents nodes as horizontal lines, and edges as vertical lines. There is one node per row, and one edge per column. To show the connection between nodes, the edges (vertical lines) are drawn darker than nodes. The ends of a node-edge link are displayed as a small square, and a looping ordered set of 32 colors are used to make tracking elements within the BioFabric over long distances.
Edges never intersect in a BioFabric visualization, making it clear which nodes are connected. Paired with the coloring system, this makes it easy to view the relationships that nodes with many edges have with other nodes. At the macro level, BioFabric easily shows the number of edges associated with a given node. Zooming in, BioFabric allows the user to easily trace the connections between nodes and edges by following the colored horizontal and vertical lines.
When representing a tree with many nodes and edges, the BioFabric is unable to display descriptive data relevant to the visualization. Therefore, a zoom function is necessary in order to read what each node and edge represent. Additionally, with very large datasets, it becomes difficult to differentiate rows, making it hard to tell where one row stops and the row below it begins. This makes the zoom function vital to the implementation of the BioFabric, otherwise it is impossible to read the data it is representing. Despite the issues caused by massive datasets, BioFabric still handles large numbers of nodes and edges better than the traditional node-link diagram, and effectively untangles the "hairball" that they create.
William Longabaugh has an excellent example of how BioFabric works at his website http://www.biofabric.org/gallery/pages/SuperQuickBioFabric.html. Additionally, there is a downloadable version of BioFabric written in Java that is available, which allows the user to import data and utilize basic zooming functions on the visualization. The download can be found at Longabaugh's website.
