Hudson Phillip S, Woodcock H Lee, Boresch Stefan
Department of Chemistry, University of South Florida , 4202 East Fowler Avenue, CHE205, Tampa, Florida 33620-5250, United States.
Department of Computational Biological Chemistry, Faculty of Chemistry, University of Vienna , Währingerstraße 17, A-1090 Vienna, Austria.
J Phys Chem Lett. 2015 Dec 3;6(23):4850-6. doi: 10.1021/acs.jpclett.5b02164. Epub 2015 Nov 24.
Carrying out free energy simulations (FES) using quantum mechanical (QM) Hamiltonians remains an attractive, albeit elusive goal. Renewed efforts in this area have focused on using "indirect" thermodynamic cycles to connect "low level" simulation results to "high level" free energies. The main obstacle to computing converged free energy results between molecular mechanical (MM) and QM (ΔA(MM→QM)), as recently demonstrated by us and others, is differences in the so-called "stiff" degrees of freedom (e.g., bond stretching) between the respective energy surfaces. Herein, we demonstrate that this problem can be efficiently circumvented using nonequilibrium work (NEW) techniques, i.e., Jarzynski's and Crooks' equations. Initial applications of computing ΔA(NEW)(MM→QM), for blocked amino acids alanine and serine as well as to generate butane's potentials of mean force via the indirect QM/MM FES method, showed marked improvement over traditional FES approaches.
使用量子力学(QM)哈密顿量进行自由能模拟(FES)仍然是一个有吸引力但难以实现的目标。该领域的新努力集中在使用“间接”热力学循环将“低水平”模拟结果与“高水平”自由能联系起来。正如我们和其他人最近所证明的,计算分子力学(MM)和QM之间收敛的自由能结果(ΔA(MM→QM))的主要障碍是各自能量表面之间所谓“刚性”自由度(例如键拉伸)的差异。在此,我们证明可以使用非平衡功(NEW)技术,即雅津斯基方程和克鲁克斯方程,有效地规避这个问题。计算ΔA(NEW)(MM→QM)对于受阻氨基酸丙氨酸和丝氨酸的初步应用,以及通过间接QM/MM FES方法生成丁烷的平均力势,显示出相对于传统FES方法有显著改进。