CPM Seminar

Computational evolution of biological networks and application to patterning

Paul François

Center for studies in Physics and Biology
The Rockefeller University

We propose a general algorithm to design genetic networks by evolution “in silico”. This procedure first aims at designing motifs performing a given function in a single cell such as bistable switches and oscillators without any constraint on their structure. The network topologies found are very similar to well-known genetic modules, such as circadian oscillators.

In superior animals, segmentation is the patterning process that leads to the formation of body metameric units such as segments in insects or vertebrae in vertebrates. A simpler computational model is then used to explore evolutionary pathways leading to possible segmentation networks. When the evolution is guided by a generic fitness function that just counts the number of segment boundaries, we invariably observe a very constrained evolutionary path. Surprisingly, the system spontaneously evolves towards a genetic network that implements the phenomenological “clock and wavefront” model proposed by Cooke and Zeeman in 1976. Simulations therefore propose explicit genetic interactions that may be later checked experimentally. Our computation illustrates how complex traits can evolve by the incremental addition of new functions on top of preexisting traits.