Online

Garcia-Bernardo, Javier

Computer Science

2015

MS

Populations of cells live in uncertain environments, where they encounter large variations in nutrients, oxygen and toxic compounds. In the fluctuating environment, cells can sense their surroundings and express proteins to protect themselves against harmful substances. However, if the stressor appears infrequently or abruptly, sensing can be too costly or too slow, and cells cannot rely solely on it. To hedge against the sudden appearance of a stressor, cell populations can also rely on phenotypic diversification through bet-hedging. In bet-hedging, cells exploit noise in gene expression or use multistable genetic networks to produce an heterogeneous distribution of resistance-conferring protein levels. In this thesis, we analyze novel roles of noise in biological systems. Through a combination of modeling and stochastic simulations, we find that noise can coordinate multi-component stress response mechanisms in a subset of the population with no extra cost. In addition, we use evolutionary algorithms to analyze the conditions where the benefits provided by noise in gene expression are equivalent to those of a more complicated, bistable distribution of protein levels. Our results show that for cells living in noisy fluctuating environments, both noise in gene expression and bistability show similar growth rates, meaning that noise in gene expression can be an effective bet-hedging strategy.