Statistical methods for analysing gene expression microarray data

flag

Klin Onkol 2006; 19(Suppl2): 365-368.

Summary
Last decade led to massive progress in the molecular biology methods which was accompanied by the production of large amount of data. This is also the case of the gene expression microarray technology that makes it feasible to study thousands of genes simultaneously. However, for relevant medical inference there is the need for appropriate evaluation and interpretation of this large quantity of experimental data.
This paper is dedicated to statistical methods that can be used for the evaluation of gene expression data. These methods can be split into three main categories: clustering methods, methods for identification of differentially expressed genes and classification techniques. Clustering methods can be used for finding of homogenous groups of patients or genes with similar expression profile, methods for identification of differentially expressed genes find genes specific for activity of certain biological tissue while classification techniques are used for setting up a discrimination rule for precise diagnostics of newly diagnosed patients to one of the previously defined classes.

Full text in PDF