xSyn: A Software Tool for Identifying Sophisticated 3-Way Interactions From Cancer Expression Data

BACKGROUND Constructing gene co-expression networks from cancer expression data is important for investigating the genetic mechanisms underlying cancer. However, correlation coefficients or linear regression models are not able to model sophisticated relationships among gene expression profiles. Here, we address the 3-way interaction that 2 genes' expression levels are clustered in different space locations under the control of a third gene's expression levels. RESULTS We present xSyn, a software tool for identifying such 3-way interactions from cancer gene expression data based on an optimization procedure involving the usage of UPGMA (Unweighted Pair Group Method with Arithmetic Mean) and synergy. The effectiveness is demonstrated by application to 2 real gene expression data sets. CONCLUSIONS xSyn is a useful tool for decoding the complex relationships among gene expression profiles. xSyn is available at http://www.bdxconsult.com/xSyn.html.