Prediction of Protein Loop Conformations Using the AGBNP Implicit Solvent Model and Torsion Angle Sampling
- 최초 등록일
- 2009.09.14
- 최종 저작일
- 2008.01
- 17페이지/ 한컴오피스
- 가격 3,000원
소개글
2008년 발표된
논문 summary
목차
Abstract
1. Introduction
2. Methods
3. Results
4. Discussion
5. Conclusion
본문내용
5. Conclusion
We have conducted loop conformation prediction tests on challenging benchmark sets consisting of 9- and 13-residue loops using the conformational search schemes built into PLOP to investigate the accuracy of the AGBNP implicit solvation model in conjuction with the OPLS-AA intramolecular force field. For a set of 57 9-residue loops investigated previously2426 we accurately predicted 88% of the loops using the OPLS-AA/AGBNP+ potential. This is a substantial improvement over the use of a distance-dependent dielectric model (63%) or SGB/NP, with (77%) or without (67%) the inclusion of crystal symmetry, as the implicit solvent model. A more substantial difference between implicit solvent models is apparent when examining the relative percentage of energy errors. AGBNP+ has the lowest percentage of energy errors at 3.5%, which is less than one-fifth as many as for SGB/NP (19.3%) and one-ninth as many as for distance-dependent dielectric (33.3%).
The fact that we have obtained high accuracy without crystal symmetry when using AGBNP+ suggests that the presence of crystal symmetry in the model is not crucial for reproducing the loop structures which have been experimentally determined via X-ray crystallography.
We expect importance sampling conformational search methods such as T-REMD to become an important complement to traditional discrete conformational search methods in cases when the number of degrees of freedom is large such as interacting loops, imperfect frameworks for loop prediction, etc.
It should be noted that the success rates we obtained likely overestimate the success rate obtainable in actual homology modeling applications because these tests were performed in the idealized case in which the frame of the protein surrounding the loop is known. Successful prediction in this idealized situation is a necessary but not sufficient requirement for the ability to predict the correct nativelike loop conformation with partial knowledge of the protein framework. We have begun to investigate cases in which the conformations of the protein side chains surrounding the loop are predicted at the same time as the loop conformation
참고 자료
없음