Fig. 3.

Scaled quantitative N-TIMP2 binding landscapes for (A) MMP-1, (B) MMP-3, and (C) MMP-14, which show the predicted effects of single amino acid substitutions on target affinity and specificity, as determined from the scaling regression equations. The color scale bars on the right of each panel range from blue, representing increases, to red, representing decreases, in the estimated affinity or specificity relative to N-TIMP2_WT. Grey indicates cases for which there was insufficient data to determine the score. This lack of data was a result of a data point being excluded from the analysis if the mutant’s frequency was below the minimum frequency threshold in more than one subpopulation, as explained in the Results section under Scaling of NGS scores by linear regression vs. K_i scores. Wild-type residues are shown as black letters inside the matrix; the substituting amino acid is shown on the X-axis; and the identity and position of the original substituted amino acid are shown in the Y-axis.