Exploration of a Holey Version of an NK Fitness Landscape

Evolution is an incredibly complex process that has been the subject of scientific study for well over a century. The complexity of evolution has made discovery by empirical studies alone challenging, as they often offer only a glimpse into much larger patterns. This has increased the importance of theoretical models to research in the field, which allow predictions to be made for empirical studies and enable these studies to be analyzed in a broader theoretical context. One set of models that have been especially important are those of fitness landscapes. These models describe the relationship between each genotype in a set and its fitness value and have been useful in understanding the mechanisms of divergence and speciation. Computationally, fitness landscapes can be produced to better represent the multidimensional nature of true biological systems. One insight of multidimensional landscapes is that they contain networks of genotypes of the same fitness, through which evolution and speciation could occur by neutral mutations alone. With the Nk model, the ruggedness, size and dimensionality of created networks can be adjusted. By imposing a fitness threshold on a quantitative trait specified by the Nk model, I am able to investigate how epistasis in the quantitative trait impacts the development and characteristics of these neutral networks of genotypes. In this thesis, I explore the properties of this novel fitness landscape model and examine how landscape features influence the structure and composition of neutral networks. I show that the neutral networks that exist in landscapes of varying levels of epistasic interaction respond very differently to increasing fitness thresholds.