Clustering causal relationships in genes expression data

Abstract

In this paper we apply a strategy to cluster gene expression data. In order to identify causal relationships among genes, we apply a pruning procedure [Chen et al., 1999] on the basis of the statistical cross-correlation function between couples of genes' time series. Finally we try to isolate genes' patterns in groups with positive causal relationships within groups and negative causal relation among groups. With this aim, we use a simple recursive clustering algorithm [Ailon et al., 2005]. © Springer-Verlag Berlin Heidelberg 2006.