Inference of logic networks from insertion and expression data

In this study, 43 tumors, that were induced by retroviral insertional mutagenesis, are profiled, resulting in a dataset for which both the initiating events (the viral integration sites) as well as the consequent expression profiles are available. We infer associations between insertion loci and gene expression profiles, while explicitly incorporating simple boolean logic, modelling multiple and parallel oncogenic pathways. We show that this results in the discovery of interesting causal associations between virally inserted loci and differentially expressed genes in tumorigenesis.