Cells grow, divide, differentiate, and respond to changing environments by means of an intricate regulatory program. The genetic circuitry carrying out this program is staggeringly complex yet capable of remarkable evolutionary modification for different physiological contexts and ecological niches. Thus there is a tradeoff between the seemingly incompatible objectives of complexity and evolvability. To both evolve efficiently and have robust function, the network must allow just enough variability on which selection can act while preserving its intricacy.; In this thesis, I study this interplay between networks, phenotype and evolution via a combination of concept, theory and experiment. Using genomic, phylogenetic, and expression data along with 30 years of literature from experimental molecular biology and genetics, I trace the evolution of four stress response networks (chemotaxis, spore formation, DNA uptake, and sigmaB general stress response) across several hundred diverse bacterial and archaeal lineages. At the conceptual level, I demonstrate that genes in these networks group into surprisingly well-defined evolutionary modules with distinctive rates of evolution and conserved patterns of gene expression. In many cases, the evolutionary module is also a module of defined dynamic control in the network, and differences in module from organism to organism seem to reflect niche adaptation.; With this finding, I attempt to refine the notion of pathway homology from simple gene homology to homology of interactions between pathway components. This leads to the development of a novel probabilistic model for estimating the phylogeny of a pathway. The model captures fine-grained dependency between the evolution of genes in the pathway using data on gene content, gene-protein and protein-protein interactions, and rates of sequence evolution. I apply this model along with a distance matrix method to estimate the phylogeny of several bacterial pathways: glycolysis and the citric acid cycle, chemotaxis, oxidative stress, and the general stress response. I compare these results to known phylogenies for these pathways, and discuss the predictive power of the model.; In parallel, I analyse genomic and expression data to investigate whether conserved phylogenetic distribution also implies conserved network dynamics, and vice versa. Preliminary results show that genes sharing expression patterns (transcriptional modules) are not always evolutionarily conserved, but that genes sharing phylogenetic patterns (evolutionary modules) are typically co-expressed. This finding sheds light on the conservation of a pathway in terms of its temporal and ecological roles, and provides an early insight into the evolution of pathway dynamics.; These comparative approaches begin the attempt towards a system-level understanding of how evolution is linked to the design and dynamics of regulatory networks, and how these networks function across timescales from microseconds to eons.
展开▼
