Suppr超能文献

QM/MM 方法比较以获得配体结合自由能。

Comparison of QM/MM Methods To Obtain Ligand-Binding Free Energies.

机构信息

Department of Theoretical Chemistry, Lund University , Chemical Centre, P.O. Box 124, SE-221 00 Lund, Sweden.

出版信息

J Chem Theory Comput. 2017 May 9;13(5):2245-2253. doi: 10.1021/acs.jctc.6b01217. Epub 2017 Apr 5.

Abstract

We have compared two approaches to calculate relative binding free energies employing molecular dynamics simulations at the combined quantum-mechanical/molecular mechanics (QM/MM) level. As a test case, we study the binding of nine cyclic carboxylate ligands to the octa-acid deep-cavitand host system. The ligand is treated with the semiempirical PM6-DH+ QM method. In the first approach, we perform direct alchemical QM/MM free energy perturbation (FEP). In the second, reference-potential approach, we convert the ligands with FEP at the molecular mechanics (MM) level and then perform also MM → QM/MM FEP for each ligand. We show that the two approaches give identical results within statistical uncertainty. For the reference-potential approach, the MM → QM/MM perturbation converges in terms of energies, uncertainties, and overlap measures with two intermediate states, giving a precision of 0.5-0.9 kJ/mol for all eight transformations considered. On the other hand, the QM/MM-FEP approach requires 17-18 intermediate states, showing that the reference-potential approach is more effective. Previous calculations with single-step exponential averaging (i.e., entirely avoiding QM/MM simulations) required fewer QM/MM energy calculations, but they gave worse precision and involved approximations with an unclear effect on the results.

摘要

我们比较了两种方法,通过在量子力学/分子力学(QM/MM)水平的分子动力学模拟来计算相对结合自由能。作为测试案例,我们研究了九个环状羧酸配体与八元深穴主体系统的结合。配体采用半经验 PM6-DH+QM 方法处理。在第一种方法中,我们进行直接的量子力学/分子力学自由能微扰(FEP)。在第二种,参考势方法中,我们在分子力学(MM)水平上用 FEP 转换配体,然后对每个配体也进行 MM→QM/MM FEP。我们表明,这两种方法在统计不确定性内给出了相同的结果。对于参考势方法,MM→QM/MM 微扰在能量、不确定性和重叠度量方面与两个中间状态收敛,对于考虑的所有八个变换,精度为 0.5-0.9 kJ/mol。另一方面,QM/MM-FEP 方法需要 17-18 个中间状态,表明参考势方法更有效。以前使用单步指数平均(即完全避免 QM/MM 模拟)的计算需要更少的 QM/MM 能量计算,但它们的精度较差,并且涉及到对结果影响不明确的近似。

文献AI研究员

20分钟写一篇综述,助力文献阅读效率提升50倍。

立即体验

用中文搜PubMed

大模型驱动的PubMed中文搜索引擎

马上搜索

文档翻译

学术文献翻译模型,支持多种主流文档格式。

立即体验