“…Discrepancies in both global minima of free energy surfaces and detailed local structures, such as salt bridge formation and helical content, have been observed, indicating that room exists for further improvement of the GB implicit solvent models (see Section II, C for more detailed discussions). Nevertheless, successful applications of implicit solvent models to address more specific problems in protein folding have been reported (Karanicolas and Brooks, 2004;Ohkubo and Brooks, 2003). In addition, recent years have seen quite a few successful applications of various GB implicit solvent models to ab initio structure prediction for a number of miniproteins, such as protein A ( Jang et al, 2003a), villin headpiece ( Jang et al, 2003a;Liu and Beveridge, 2002), Trp-cage (Pitera and Swope, 2003), Trp-zip (Okur et al, 2003;Steinbach, 2004;Yang et al, 2004), bba motifs (Jang et al, 2003b), and the fd Coat transmembrane protein .…”