How predictable is regulatory network evolution?

Gábor Balázsi

Time: 9:30 - 10:10 AM

Type: Invited presentation

Affiliation: Laufer Center for Physical and Quantitative Biology,
Biomedical Engineering Department, Stony Brook University

Abstract

The evolution of gene regulatory networks is poorly understood, partly because we lack appropriate model systems that allow the development of quantitative, experimentally testable predictions. To address this problem, we inserted a synthetic gene network module into the budding yeast genome and followed its evolution in various environments that affected its costs and benefits to the host. In agreement with computational predictions, we found that mutations: (i) target and eliminate the module if it has only cost; (ii) activate the module if it is potentially beneficial and carries no cost; and (iii) fine-tune the module's response if it has excessive cost and/or insufficient benefit. These results suggest that gene network evolution may be predictable from the interplay of environment-dependent costs and benefits with network dynamics, all of which in principle can be determined prior to evolution.