Boolean Networks : a Study on Microarray Data Discretization

Abstract

Biomedical research has been revolutionized by high-throughput techniques and the enor mous amount of biological data they are able to generate. Genetic networks arise as an es sential task to mine these data since they ex plain the function of genes in terms of how they influence other genes. Genetic networks based on discrete states, such as boolean net works, have been widely used and have shown abilities to model some of the complex dy namics of gene expression networks. In this work we propose a new method for the dis cretization of gene expression data based on the fuzzification of already proposed tech niques. The proposal is applied to the mi croarray data obtained from a problem on the inflammation and host response to injury in human beings.