The Population Genetics of Biological Noise

Information transmission is intrinsic to life, and noise is intrinsic to information transmission. Biological noise during development is essential for the flexibility and plasticity of individual organisms, but also underlies some diseases. Biological noise during reproduction is the fuel for evolution, including the evolution of therapy resistance in pathogenic microbes and in cancer. Recent technological advances in our ability to characterize many sources of biological noise have demonstrated that its amount is often heritable. Here, we frame the population genetics of loci that influence the amount of any source of biological noise. While analogous theory for heritable changes in mean trait values has been established for nearly a century, to our knowledge this is the first general approach for studying the evolution of heritable changes in their statistical distributions. This represents a critical theoretical contribution to an important and rapidly growing domain of intellectual inquiry. It also sheds light on the hypothesis that natural selection can increase evolvability, and generalizes modifier theory used in the tradition of Feldman and colleagues. Author summaryBiological noise is a fact of life. Genetically identical organisms reared in identical environments invariably exhibit random phenotypic differences. And siblings born of the same parent(s) in the same environment are endowed with inheritances that invariably differ at random. While the specific consequences of biological noise are unpredictable, extraordinary experimental advances now make clear that its amount can be influenced by an organisms genetics. For example, high- and low-noise promoter, and high- and low-noise DNA polymerase alleles are well known. This raises the question of when and how natural selection favors high- or low-noise alleles. While biological noise is on average deleterious, it can also occasionally induce high fitness phenotypes. Here, we solve a simple analytic model for the fitness difference between noise alleles that captures both these features. Our model predicts the existence of an evolutionary equilibrium in the amount of noise, whose location reflects just four features of an organisms biology. In light the clinical importance of biological noise, as well as its central role more broadly in both development and evolution, this work provides an urgently needed evolutionary framework for understanding its long-term determinants.