A summer with genes : simple disease classification from microarray data.

Abstract

In this article we report on the work carried out within the framework of a summer project, part-funded by an IMA small grant, in which an undergraduate student (the second author of this manuscript) developed and implemented methodology for disease classification from gene expression microarray data. While the original motivation for this study was the development of a correlation threshold for gene filtering, a general outcome of this research was that, using very simple statistical techniques (essentially at undergraduate level) but solid state-of--the-art validation routines, good classification accuracies can be obtained using relatively small-sized gene signatures. We applied the techniques on expression data for breast cancer tumour subtype classification, as well as for prediction of the presence or absence of Irritable Bowel Syndrome (IBS).