Evolution and genotype-phenotype maps in laboratory microbial populations

Abstract

Understanding how genetic variation maps to phenotypic outcomes is central to predicting evolutionary dynamics. Evolution acts on phenotypes, yet genetic change occurs at the level of genotypes, making the genotype–phenotype (GP) map a fundamental bridge between molecular variation and fitness. This map is often complex: mutations interact nonlinearly, depend on genetic background and environment, and can produce emergent effects that are difficult to predict. In this thesis, I investigate the structure of GP maps and their consequences for adaptation, focusing on epistasis, higher-order interactions, and spatial structure in microbial populations. Using high-throughput experiments in budding yeast and theoretical modeling, I show that higher-order epistasis contributes weakly to phenotypic variation, with additive and pairwise models capturing most explainable variance, suggesting an effectively low-dimensional structure. I then develop and test theoretical models of complex traits with widespread epistasis and pleiotropy, demonstrating that simple, predictive patterns emerge and hold across genetic backgrounds and environments. Next, I examine phenotypic convergence despite genetic stochasticity in long-term evolution experiments, showing how degenerate genetics shape mutational effects. Finally, I show that spatial structure can alter evolutionary dynamics and distort genetic signatures, particularly in the site frequency spectrum. Together, these results provide a framework for understanding and predicting genotype-phenotype maps and evolutionary outcomes, in both well-mixed and spatially structured populations.