An illustration of our method for finding specificity determining residues (SDRs). The sequences of the peptide recognition domains (PRDs) are aligned to form a multiple sequence alignment (MSA). Each PRD has an associated position weight matrix (PWM) of the peptides that it binds, and the PWMs are also aligned. The entropy of each MSA site X and each PWM site Y is computed and combined to give the mutual information (MI) and uncertainty coefficient (UC) scores.
Yip et al. BMC Biology 2011 9:53 doi:10.1186/1741-7007-9-53