Methodology

The Shivplot: a graphical display for trend elucidation and exploratory analysis of microarray data

Owen Z Woody^1,2 and Robert Nadon^1,3

¹  McGill University and Genome Quebec Innovation Centre, 740 avenue du Docteur Penfield, Montreal, Quebec, H3A 1A4, Canada

²  David R. Cheriton School of Computer Science, University of Waterloo, Waterloo, Ontario, N2L 3G1, Canada

³  Department of Human Genetics, McGill University, Montreal, Quebec, Canada

author email corresponding author email

Source Code for Biology and Medicine 2006, 1:6doi:10.1186/1751-0473-1-6

Published:	8 November 2006

Abstract

Background

High-throughput systems are powerful tools for the life science research community. The complexity and volume of data from these systems, however, demand special treatment. Graphical tools are needed to evaluate many aspects of the data throughout the analysis process because plots can provide quality assessments for thousands of values simultaneously. The utility of a plot, in turn, is contingent on both its interpretability and its efficiency.

Results

The shivplot, a graphical technique motivated by microarrays but applicable to any replicated high-throughput data set, is described. The plot capitalizes on the strengths of three well-established plotting graphics – a boxplot, a distribution density plot, and a variability vs intensity plot – by effectively combining them into a single representation.

Conclusion

The utility of the new display is illustrated with microarray data sets. The proposed graph, retaining all the information of its precursors, conserves space and minimizes redundancy, but also highlights features of the data that would be difficult to appreciate from the individual display components. We recommend the use of the shivplot both for exploratory data analysis and for the communication of experimental data in publications.