Enriched Heatmaps for Visualizing Uncertainty in Microarray Data

Abstract

Heatmaps for microarray visualization represent the measured and processed signal intensities in a matrix. Each cell of the matrix represents one gene of a given sample and the corresponding data value is mapped onto a red/green color scale, with red indicating up-regulation and green indicating down-regulation. To enhance heatmaps with uncertainty information we need to map the QCS values to visual attributes that can be incorporated into the heatmap without hampering its perception. Mapping uncertainty to free visual attributes such as brightness and saturation bears the risk of misinterpretation of a cell's color, while other attributes such as orientation or the cell size potentially interfere with neighboring cells.Our approach proposes a deformation of the heatmap's grid lines that equates a mapping of the QCS values onto the size of an additional geometric shape within each cell. This does not interfere with the perception of a cell's color or its frame within the matrix.