Fluidity and Predictability of Epistasis on an Intragenic Fitness Landscape

How epistasis hinders or facilitates movement on fitness landscapes has been a longstanding question in evolutionary biology. High-throughput experiments have revealed that, despite their idiosyncratic nature, epistatic interactions can exhibit reproducible global statistical patterns. Recently, Papkou et al. constructed a fitness landscape for a 9-base pair region of the folA gene in Escherichia coli , which encodes dihydrofolate reductase (DHFR), and showed that this landscape is both rugged and highly navigable. Here, we analyze this landscape to address two questions: How does the nature of epistasis between two mutations change with genetic background? and How predictable is epistasis within a gene? We find that epistasis is “fluid”: higher-order interactions cause the relationship between two mutations to shift strongly across genetic backgrounds. Mutations fall into two distinct categories: a small subset exhibit strong global epistasis, while the majority do not. Nonetheless, we find that the distribution of fitness effects (DFE) of a genotype is highly predictable from its fitness. These findings provide a gene-level perspective on how epistasis operates, revealing both its unpredictability and its statistical regularities, and offer a framework for predicting mutational effects from high-dimensional fitness landscapes.