Getting poster data...
Julian Heinrich, Corinna Vehlow, Kay Nieselt, Florian Battke, Daniel Weiskopf (Visualization Research Center (VISUS), Stuttgart, Germany; Integrative Transcriptomics, Tübingen, Germany)We present iHAT, the interactive hierarchical aggregation table. Similar to the heatmap, iHAT consists of a table linked with a dendrogram. Key difference to other tools is the support for interactive construction of trees from arbitrary data. Rows selected by the user are aggregated into new internal nodes. Depending on the type of data per column, different aggregation methods are used: Metric data is aggregated using the mean, nominal data using the consensus of the selected rows. For metric columns, we use a single-hue colormap with varying saturation to indicate the corresponding value. For nominal columns, we adapt the number of different hues to the number of classes contained in the respective column and map saturation to the frequency of the consensus to visualize its uncertainty. We found that iHAT can be useful for the exploration of genomic data, especially if metadata such as phenotype information is available. Starting from a multiple sequence alignment and using metadata to guide the aggregation of rows, we were able to reveal correlations between SNP and phenotypic patterns in a first test using artificial data.