(A) The relationship between the average order parameter in distant, non-solvent-exposed residues versus the average order parameters in binding site residues (n = 743, slope = 0.79, r2 = 0.65; p=6.5 × 10–89, two-sided t-test). (B) We compare the difference in order parameters in each binding site residues of holo-apo pairs compared to a control dataset made up of the same number, type, and solvent exposure of amino acids. Comparing the apo/holo structures, on average binding site residues got more rigid upon binding. The median difference in order parameters was 0.03 for the binding site residues compared to 0 for the control dataset (p=3.4 × 10–7, individual Mann–Whitney U test). (C) The relationship between the residual order parameters in all distant residues versus binding site residue order parameters (n = 743, slope = −0.34, r2 = 0.17; p=4.6 × 10–28, two-sided t-test). (D) The relationship between the residual order parameters in distant, non-solvent-exposed residues versus binding site residues in the apo and apo control dataset residues (n = 283, slope = −0.28, r2 = 0.20; p=1.8 × 10–34, two-sided t-test). (E) To analyze the impact that protein size has on the relationship between the difference in order parameters in binding site residues versus the residual order parameters in distant residues, we binned proteins based on the number of residues. We colored each point on our original binding site residues versus the residual order parameters in distant residues based on the protein size and did not observe any clustering. (F) The bootstrap analysis of the overlap of the slope of distant, average order parameters of non-solvent-exposed residue versus average order parameters of binding site residue between holo-apo (green) and apo-apo (purple). While there was some overlap, the mean slope of holo-apo (–0.44) was more than 2 standard deviations away from the mean slope of the apo-apo (–0.28). Comparing the two bootstrap distributions using a z-test, the z-value was –191.26 with a p-value of 0.0.