An End to Endless Forms: Epistasis, Phenotype Distribution Bias, and Nonuniform Evolution
(A) The average Hamming distance among visible phenotypes as a function of their frequency (dots). Visible phenotypes are ranked according to their frequency level. For each rank, we calculate the average Hamming distance between all visible phenotypes with this or higher rank. The most abundant phenotypes are very similar. This similarity decreases as less frequent phenotypes are included in the analysis. We also calculate which fraction of all visible phenotypes are included in these phenotypes (solid line). The inset shows a zoom of the same plot, focusing only on the top 5% most frequent phenotypes. The phenotypes that are included in this small fraction of the distinct visible phenotypes, are, on average, only 4 bits different, and still cover 50% of the phenotypes. (B) The one mutant neighbor network of the visible phenotypes. The size of the node is proportional to the logarithm of its frequency. In this plot, r = k = 12.