VisGenome with CartoonPlus: Supporting large scale genomic analyses via physical space deformation

Abstract

We focus on visualisation techniques used in genome browsers and examine the available browsers with respect to their suitability for comparative genome analysis, and the legibility of display. Based on this investigation, we then report on a new technique, CartoonPlus, which improves the visual representation of data. We describe our use of smooth zooming and panning, and a new scaling algorithm and focus on options. CartoonPlus scaling allows the users to see data not in a physical size, but deformed to improve data legibility, depending on the data type chosen by the users. In VisGenome we have chosen genes as the basis for scaling. All genes have the same size and all other data is scaled in relationship to genes. Additionally, objects which are smaller than genes, such as micro array probes or markers, are scaled differently to reflect their partitioning into two categories: objects in a gene region and objects positioned between genes. This results in a significant legibility improvement and should enhance the understanding of genome maps. The technique may be useful in other information rich contexts, such as comparison of histograms or schema mapping.