Lack of evidence for sign epistasis between beneficial mutations in an RNA bacteriophage

In previous work (Betancourt, Genetics 181:1535, 2009), I propagated three large laboratory populations of an RNA phage (MS2) as they adapted to a controlled laboratory environment. These populations were large enough so that evolution might be expected to be mostly repeatable, but they nevertheless fixed different suites of mutations over the course of the experiment. Here, I investigate one possible explanation for these results: epistasis, in which the effect of a mutation depends on its genetic background, may have prevented populations with different initial substitutions from fixing the same set of subsequent mutations. I show that two mutations that previously occurred in different genetic backgrounds are beneficial on either background. This result suggests that sign epistasis-in which a mutation is beneficial on one background, but deleterious on another-is not the cause of different evolutionary trajectories observed in the Betancourt (2009) experiment. However, they can be explained by either magnitude epistasis-in which mutations have stronger or weaker beneficial effects depending on the background-or by the simultaneous fixation of multiple beneficial mutations. Surprisingly, the large populations of the previous experiment showed less parallel evolution than the small populations of this experiment, which lends support to the fixation of multiple beneficial mutations contributing to the patterns seen in both experiments.