Discrete causal model view of biological networks

Abstract

Signaling and regulatory pathways coordinate multiple cellular functions in response to environmental variations. Discovering the pathways governing functionally specific responses is essential for understanding of biological systems. It aims at determining the causal cascades of regulations leading to the observed responses. Their characterization by computational methods remains an important and challenging question. The presented cabin (Causal Analysis of Biological Interaction Network) method determines a causal model view composed of a subnetwork and a set of agent states deduced from observations with regards to a model of network dynamics. The validity of the results is ensured by formally checking the conditions of correctness of a model with respect to observations. State-based and symbolic versions of the algorithm have been implemented and used for a biological case study.