Fig. 7: The collective influence of the set of mutations in each design on the predicted effect of individual point mutations.

[A, left] Example design (70.a) of the raw data that are summarized in the table on the right. Plotted is a comparison of how the addition of each point mutation in 70.a changes the predicted fitness (ΔEVH) in WT TEM-1 (x-axis) versus 70.a (y-axis). Each of the 88 points represent a single amino acid substitution that is found in 70.a. Pink background indicates the top left quadrant, which are those mutations that are predicted to have a negative effect on fitness in WT TEM-1 and positive effect on fitness in 70.a. The percent of amino acid changes in each quadrant is indicated. [A, right] Summary of predicted fitness changes (ΔEVH) for each designs' amino acid changes compared with their predicted effect in WT TEM-1. The “percent of mutations” is the percent of mutations in each quadrant (see example on left), with the pink background containing the percentage that are predicted to have a negative fitness effect in WT TEM-1 and positive effect in the design. B Distribution of predicted fitness effects (ΔEVH) of all possible amino acid changes (252 * 19) introduced in the WT TEM-1 sequence background (top, blue) and 70.a sequence background (bottom, blue). Red lines are the predicted fitness effect of the 70.a mutations if added individually to WT TEM-1 (top) or removed from 70.a (i.e., reverted to the WT TEM-1 amino acid). Source data are provided in the Source Data file.