Protein tertiary structure recognition using optimized associative memory Hamiltonians

Protein folding codes can be optimized using spin-glass theory. Simple optimal codes are deduced for associative memory Hamiltonians based on aligned sequences. The authors demonstrate the use of associate memory Hamiltonians and show how, with the use of spin-glass theory to provide optimization, and standard alignment methods, it is possible to generate an algorithm that correctly identifies structurally similar proteins in the vast majority of cases.