The importance of epistasis in the evolution of viral pathogens

Understanding the genetic and genomic underpinnings of infectious disease outbreaks has emerged as a frontier of epidemiology. Here, we argue that epistasis-where the phenotypic effects of mutations or gene variants are dictated by the presence of other mutations or genes-should become a focus of genomic epidemiology. To demonstrate this, we present the results of a systematic review of the literature on epistasis in viruses, focusing on three major human viral systems: (i) influenza, (ii) SARS-CoV-2, and (iii) human immunodeficiency virus, as well as two other bodies of the literature mainly focusing on nonhuman viruses: (iv) tobacco etch virus and (v) experimental evolution of viruses. Our systematic review of these five bodies of the literature highlights that epistasis is prevalent in host-virus systems of various kinds, manifesting within and between different loci, with effects of different magnitudes and directions, and shaping various phenotypic traits of epidemiological interest. At the same time, our systematic review demonstrates that our ability to draw general conclusions about the direction and magnitude of epistasis in viral evolution is constrained by several factors: the idiosyncrasies of virus-host systems, biases in the underlying data collection exercises, and the limitations of existing methods. Moving forward, we encourage collaborations between genomic epidemiologists and evolutionary biologists to identify and measure epistasis in studying the evolution of viral pathogens.