Evolution of Recombination on an HIV-1 Derived Fitness Landscape

The maintenance of recombination is among the most important unsolved problems of evolutionary biology. The Hill-Robertson effect, which states that the interaction between genetic drift and selection generates unfavorable linkage dise-quilibria (hence favoring recombination), offers one of the most promising hypotheses to solve this problem. In particular,it has been argued that this hypothesis works independently of epistatic interactions. However, this result has been derivedon the basis of smooth fitness landscapes, which may be unrealistic (Otto and Feldman (1997)). We estimated the fitnesseffects of 1’857 single mutations and of 257’536 pairs of mutations found in a 60’000 HIV-1 B pol-genotypes assayed forin vitro replication capacity (Hinkley et al. (2010)) to develop a reasonably realistic model of a fitness landscape on whichwe run a genetic algorithm to mimic the evolution of HIV populations. By adding a recombination rate modifier to thegenome, we address the question of whether genetic drift outweighs epistasis as a factor for the evolutionary maintenanceof recombination in the case of the fitness landscape of our model. Despite the fairly rugged nature of the fitness landscape,which could be characterized by the presence of a large number of local optima, we find that recombination is robustlyfavored in finite populations. This result suggests that the Hill-Robertson effect provides a powerful explanation for theevolutionary maintenance of recombination even if fitness landscapes are rugged. ReferencesOtto, S. and Feldman, M., (1997). Deleterious mutations, variable epistatic interactions, and the evolution of recombination.Theor Popul Biol 51: 134–147. Hinkley, T., Petropoulos, C., and Bonhoeffer, S. (2010). The Evolutionary Systems Biology of HIV Drug Resistance. In press.