Table 1.

Comparison of computational protocols to experimental data

Computational method	Jensen–Shannon divergence	Euclidean distance
Standard	0.472	0.839
DNA-Rebuild	0.409	0.670
HighTemp-Packer	0.399	0.695

Divergence between experimentally observed and computationally predicted amino acid frequency distributions at 44 positions of the I-AniI protein–DNA interface was assessed using two standard metrics for comparing probability distributions: the Jensen–Shannon divergence³³ and the Euclidean distance. A lower divergence value indicates that the probability distributions better match one another.