Pocket Similarity: Are α Carbons Enough?

摘要: A novel method for measuring protein pocket similarity was devised, using only the α carbon positions of residues. Pockets were compared pairwise an exhaustive three-dimensional Cα common subset search, grouping residues by physicochemical properties. At least five matches required each hit, and distances between corresponding points fit to Extreme Value Distribution resulting in a probabilistic score or likelihood any given superposition. set 85 structures from 13 diverse families clustered based on binding sites alone, this score. It also successfully used cluster 25 kinases into number subfamilies. Using test kinase query retrieve other pockets, it found that specificity 99.2% sensitivity 97.5% could be achieved appropriate cutoff The search itself took 2 10 min single 3.4 GHz CPU entire Protein Data Bank (133 800 pockets), depending hits re...