VistaClara

VistaClara is created to explre the effectiveness of applying an extended permutation matrix to the task of exploratory data analysis of multi-experiment microarray studies, and provides a dynamic compressed heatmap view of the entire data set that makes it possible to perform highly interactive ad hoc visual explorations of microarray.

Abstract

We have created VistaClara to explre the effectiveness of applying an extended permutation matrix to the task of exploratory data analysis of multi-experiment microarray studies. The permutation matrix is a visualization technique for interactive exploratory analysis of tabular data that permits both row and column rearrangement, and fits well with the tabular forms of data characteristic of gene expression studies. However, this technique has been largely overlooked by current bioinformatics research. Our implementation supports direct incorporation of supplemental data and annotations into the matrix view. This enables visually searching for patterns in gene expression measurements that correlate with other types of relevant data (disease classes, clinical, histological, drug treatments, etc.). The heatmap visualization common in microarray analysis is extended to provide a novel alternative using size as well as color to graphically represent experimental values, thus allowing more effective quantitative comparisons. Methods to sort rows or columns by similarity extend the possible permutation operations, and allow more efficient searching for biologically relevant patterns in very large data sets. Based on overview+detail principles, a dynamic compressed heatmap view of the entire data set provides the user with overall context, including possible correlations not currently visible in the more detailed view. Combined, these techniques make it possible to perform highly interactive ad hoc visual explorations of microarray.