Source:
Computational Methods for Protein Folding, 120, 2002Abstract
The major focus of this review is to describe a unified approach to protein structure prediction that reduces to threading plus structure refinement when an example of the probe sequence is found; but if not, it incorporates information from weakly significant probe sequence-template structure matches and then does ab initio folding with the structural information gleaned from such matches. It has the advantage that it can predict a novel fold even though some of the information comes from threading on already solved structures.