Graphically Displaying Partitions of Variation in Microarray Data David L. Gold , The University of Texas, Houston Two-way Hierarchical Clustering Plots are a widely accepted method for displaying the partitions of variation in microarray data in the bioinformatics literature. Although hierarchical clustering methods are not new, few methods have been explored to partition multivariate array data sets in both the column and row space. A graphical approach is offered which allows the user control in displaying variation ordered in either or both the array sample and gene space. Fixed and random sources of variation may be defined explicitly in the algorithm. In addition a technique is offered to identify dominant source(s) of variation across samples. |
Graybill Conference |