Mathematics Colloquium: Testing Assumptions of Models of Adaptations Using Experimental Evolution
4:10 pm Webs 11
Much work has been done to modeling the adaptive process in haploid organisms, but comparatively little data has been gathered testing the validity of model assumptions. To this end, we conducted a flask-passage adaptation experiment at two bottleneck sizes (104 and 106) on a G4-like bacteriophage (ID11) followed by extensive whole-genome sequencing and fitness assays. The dataset provides a brief, but detailed look at the adaptive dynamics. The small bottleneck lines appear near the transition between selective sweep conditions and those that generate more complex interference dynamics (eg. clonal interference). The large bottleneck lines exhibit extensive interference, less stochasticity, and have multiple beneficial mutations establishing on a variety of backgrounds. The distribution of beneficial effects of mutations on both the wildtype background and on a first-step background failed to follow an exponential distribution, but instead supported a bounded distribution. The boundary of the distributions differed, however, between backgrounds. This, along with a large number of beneficial mutations observed on a highly fit first-step background, suggests the fitness landscape is not fully uncorrelated. Nor does the data, with a rarity of first-step beneficial mutations appearing as second-steps, support an additive, smooth fitness landscape. Taken together, the data point to incongruities between the biology and some common modeling assumptions, however, a version of Fisherâ€™s geometric model provides some interesting insights.