Complementary use of cluster analysis and biplots to discover and validate patterns of gene expression in microarray data

Microarray studies are used in molecular biology to explore patterns of expression of thousands of genes. This methodology has relevantly developed in the last decades, and so has the need for appropriate methods for analyzing high-throughput data generated from such experiments. Identifying sets of genes and samples characterized by similar values of expression and validating these results are two of the issues related to these investigations. From a statistical perspective there is no general agreement on these problems. Specifically, the use of Cluster Analysis is often a critical relying on the main use of hierarchical techniques without considering possible use of other methods. Moreover, validation of results using external datasets is still subject of discussion. In this paper we show the use of several clustering algorithms to discover common patterns of expression, and propose a rank-based passive projection of Principal Components for validation purposes. Results from a study involving 23 cell lines and 76 genes are presented.