Using multiagent system for gene expression classification

Recent advances in microarray technology offer the ability to measure expression levels of thousands of genes simultaneously. Analysis of such data can help us identifying different clinical outcomes using only expressions of a few predictive genes. This paper presents an application of multiagent system to the analysis of gene expression data. Our goal is to find significant classification genes using simple classifiers that can be used by agents when exploring the gene expression database. We present our results on two well-known publicly available gene expression problems where we try to achieve the highest possible accuracy of classification using the smallest possible set of genes.