A new substitution matrix for protein sequence searches based on contact frequencies in protein structures

摘要: The instabilities of the native structures mutant proteins with an amino acid exchange are estimated by using contact energy and number contacts for each type pair, which were from 18,192 residue-residue observed in 42 crystals globular proteins. They then used to evaluate a transition probability matrix codon substitutions log relatedness odds matrix, is as scoring measure similarity between protein sequences. To consider homologous proteins, base mutation rates effects genetic code also taken into account. average fitness approximated be proportional structural stability protein, change structure expected neglect denatured state. In global local homology searches, this tends yield significantly higher alignment scores than either unitary or may distantly related pairs MDM78. One advantages that equilibrium frequencies codons can adjusted.