Novel approaches to biclustering and gene functional classification in microarray gene expression data

Abstract

Microarray analysis is a high-throughput experimental technique with the capacity to measure the expressions of thousands of genes in parallel over many experimental samples (tissues types, environmental conditions, time points etc.). To fully exploit the large volumes of expression data produced by these experiments requires the application of statistical analysis and machine learning methods. Microarray datasets may contain many genes and samples with unknown labels. New gene functional classes may also emerge as our understanding of the underlying biological system increases. As a result, unsupervised methods of analysis, such ais cluster analysis, often prove most useful in this domain.