Systems biology Advance Access publication January 8, 2014 Characterizing cancer subtypes as attractors of

Motivation: Cancer is a heterogeneous progressive disease caused by perturbations of the underlying gene regulatory network that can be described by dynamic models. These dynamics are commonly mod-eled as Boolean networks or as ordinary differential equations. Their inference from data is computationally challenging, and at least partial knowledge of the regulatory network and its kinetic parameters is usually required to construct predictive models. Results: Here, we construct Hopfield networks from static gene-expression data and demonstrate that cancer subtypes can be char-acterized by different attractors of the Hopfield network. We evaluate the clustering performance of the network and find that it is compar-able with traditional methods but offers additional advantages includ-ing a dynamic model of the energy landscape and a unification of clustering, feature selection and network inference. We visualize the Hopfield attractor landscape and propose a pruning method to gen-erate sparse networks for feature selection and improved understand-ing of feature relationships. Availability: Software and datasets are available a