The genome-wide deleterious mutation rate in humans has a lower bound of Ud>2.1. However, detailed balance between deleterious and beneficial alleles is impossible for Ud>1 (Kondrashov 1995). Instead of equilibrium, a smaller number of large effect beneficial fixations compensate for a larger number of small effect deleterious mutations. This can lead to ratcheting complexity in a similar way to the drift barrier hypothesis, but differs in that it requires high Ud but does not require low population size N (Matheson et al. 2023). With high Ud, neutral genetic diversity is depressed more by background selection from unlinked sites than from linked sites (Matheson & Masel 2023). Widespread linkage disequilibrium between unlinked sites can be seen in simulations, as low variance in the number of deleterious mutations per individual, below the mean. This pattern was previously documented empirically, but interpreted as evidence for synergistic epistasis (Sohail et al. 2017). We show it arises due to background selection alone in the non-equilibrium Ud>1 regime. Our finding of such strong background selection also has implications for population genetics in the wake of population bottlenecks, which sample rare deleterious mutations. When recessive, this intensifies already-strong background selection, in populations that are already beyond the limit of purging.
High deleterious mutation rates drive non-equilibrium dynamics