Critical Assessment of Self-Consistency Checks in the All-Atom Molecular Dynamics Simulation of Intrinsically Disordered Proteins.
Critical Assessment of Self-Consistency Checks in the All-Atom Molecular Dynamics Simulation of Intrinsically Disordered Proteins.
复制标题
DOI：
10.1021/acs.jctc.2c01140
复制
发表时间：
2023-05-23
影响因子：
5.5
通讯作者：
Ghosh, Kingshuk
中科院分区：
文献类型：
--
作者：
Gaalswyk, Kari;Haider, Austin;Ghosh, Kingshuk
所有的原子模拟都可以用来定量分析蛋白质的构象特性。然而，仿真必须满足收敛检查，以确保从仿真计算的观测值是可靠的和可再现的。虽然绝对收敛纯粹是一个理论概念，需要无限长的模拟，但更实际，但严格的方法是施加自一致性检查（SCC），以获得模拟数据的信心。目前没有对国内流离失所者中的SCC进行研究，这与折叠的SCC不同。在本文中，我们介绍了不同的标准，为国内流离失所者的自我一致性检查。接下来，我们施加这些SCC严格评估不同的模拟方案的性能，使用N末端结构域的HIV整合酶和连接区的SARS-CoV-2核蛋白作为两个模型IDP。所有的模拟协议开始与全原子隐式溶剂蒙特卡罗（MC）模拟和随后的MC生成的构象的聚类，以创建的IDP的代表性结构。这些代表性的结构作为初始结构的后续分子动力学（MD）运行明确的溶剂。我们的结论是，生成多个短（~3 μs）MD模拟轨迹-所有从最具代表性的MC生成的构象开始-并将它们合并是选择的协议，因为（i）它能够满足多个SCC，（ii）一致地再现实验数据，以及（iii）通过利用现代GPU集群中可用的多个核心并行运行独立轨迹的效率。运行一个长轨迹（大于20 μs）也可以满足前两个标准，但由于计算时间太长而不太理想。这些发现有助于解决识别可用起始配置的挑战，提供SCC的客观测量，并建立严格的标准来确定IDPs的全原子模拟所需的最小长度（对于一个长模拟）或轨道数量。
All atom simulations can be used to quantify conformational properties of Intrinsically Disordered Proteins (IDP). However, simulations must satisfy convergence checks to ensure observables computed from simulation are reliable and reproducible. While absolute convergence is purely a theoretical concept requiring infinitely long simulation, a more practical, yet rigorous, approach is to impose Self Consistency Checks (SCCs) to gain confidence in the simulated data. Currently there is no study of SCCs in IDPs, unlike their folded counterparts. In this paper, we introduce different criteria for self-consistency checks for IDPs. Next, we impose these SCCs to critically assess the performance of different simulation protocols using the N terminal domain of HIV Integrase and the linker region of SARS-CoV-2 Nucleoprotein as two model IDPs. All simulation protocols begin with all-atom implicit solvent Monte Carlo (MC) simulation and subsequent clustering of MC generated conformations to create the representative structures of the IDPs. These representative structures serve as the initial structure for subsequent molecular dynamics (MD) runs with explicit solvent. We conclude that generating multiple short (~3 μs) MD simulation trajectories—all starting from the most representative MC generated conformation—and merging them is the protocol of choice due to (i) its ability to satisfy multiple SCCs, (ii) consistently reproducing experimental data, and (iii) the efficiency of running independent trajectories in parallel by harnessing multiple cores available in modern GPU clusters. Running one long trajectory (greater than 20 μs) can also satisfy the first two criteria but is less desirable due to prohibitive computation time. These findings help resolve the challenge of identifying a usable starting configuration, provide an objective measure of SCC, and establish rigorous criteria to determine the minimum length (for one long simulation) or number of trajectories needed in all-atom simulation of IDPs.
登录
查看更多内容
影响因子：
6.8
作者：
Dunker, A. Keith;Silman, Israel;Sussman, Joel L.
通讯作者：
Sussman, Joel L.
影响因子：
3
作者：
Arcangeli, Caterina;Cantale, Cristina;Rosato, Vittorio
通讯作者：
Rosato, Vittorio
影响因子：
3.3
作者：
Genheden, Samuel;Ryde, Ulf
通讯作者：
Ryde, Ulf
DOI：
10.1073/pnas.1207719109
发表时间：
2012-10-02
影响因子：
11.1
作者：
Hofmann, Hagen;Soranno, Andrea;Schuler, Benjamin
通讯作者：
Schuler, Benjamin
影响因子：
2.9
作者：
Faraldo-Gómez, JD;Forrest, LR;Sansom, MSP
通讯作者：
Sansom, MSP