Genome-wide mouse embryonic stem cell regulatory network self-organisation:a big data CoSMoS computational modelling approach.

The principal barrier to gaining understanding of embryonic stem (ES) cell regulatory networks is their complexity. Reductionist approaches overlook much of the complexity inherent in these networks and treat the ES cell regulatory system as more or less equivalent to the sum of its component parts, studying them in relative isolation. However, as we learn more about regulatory components it becomes increasingly difficult to integrate complex layers of knowledge and to develop more refined understanding. We seek better control of the complexity inherent in non-equilibrium ES cell regulatory networks undergoing lineage specification by developing computer simulations of self-organisation using the CoSMoS approach. Simulation, together with the hypothesis that lineage computation occurs at the edge of chaos, should allow us to investigate the driving of gradual accumulation of network complexity 'from the bottom up'. Here, we present the first step in this design process: use of the CoSMoS approach to develop a highly abstracted model and simulation of regulatory network activity driven by just single pluripotent transcription factors (TF), but at genome-wide scales. We investigate three TFs in isolation: Oct4, Nanog and Sox2, central elements of the core pluripotent network of mouse embryonic stem cells. This provides a suitable basis for future modelling of multiple interacting TFs.