Accurate and Efficient Prediction of NMR Parameters of Condensed-Phase Systems with the Generalized Energy-Based Fragmentation Method

Abstract

We have implemented the calculations of NMR parameters within the generalized energy-based fragmentation (GEBF) method for condensed-phase systems with periodic boundary conditions (PBC). In this PBC-GEBF approach, NMR parameters of molecules in a unit cell are assembled as a linear combination of the corresponding quantities from a series of small embedded subsystems. To treat condensed-phase systems containing large molecules, we propose a novel “fragment-based” strategy for building subsystems, while our previously reported “molecule-based” strategy for construction of subsystems is appropriate for periodic systems with small molecules. The “fragment-based” strategy in PBC-GEBF is demonstrated to be much more efficient than its “molecule-based” counterpart to treat crystals of large molecules. With the “molecule-based” PBC-GEBF method, we obtained consistently good NMR parameters of liquid water with B3LYP on top of neural-network-potential-based ab initio molecular dynamics (AIMD) snapshots. With the “fragment-based” PBC-GEBF approach, we predicted the 1H chemical shifts of a large macrocycle in solution based on a series of classical MD snapshots. The calculated results are in good accord with the experimental chemical shifts. Therefore, the PBC-GEBF method is expected to be a reliable and efficient tool for predicting NMR parameters of large complex systems in solutions.

Structures and properties of ionic crystals and condensed phase ionic liquids predicted with the generalized energy‐based fragmentation method

“…In Figure 4, we present a typical snapshot structure of this studied ILs as well as the structure of the [C 2 mim] + cation (with labeled H types). Our recent work demonstrated that for liquid water and a large macrocycle in solution, the ensemble‐averaged 1 H chemical shifts (over a few MD snapshots) obtained with the PBC‐GEBF approach at the B97‐2/pcSseg‐2 level are in good agreement with the experimental results 66 . Hence, to simulate the 1 H chemical shifts of [C 2 mim][BF 4 ] ILs, some snapshots of this ILs were also equidistantly selected along the MD trajectories.…”

“…b The experimental values are from ref 73. c Deviations of the PBC-GEBF calculated values relative to the corresponding experimental values. F I G U R E 4 Snapshot of the [C 2 mim] [BF 4 ] ILs (with 20 ion pairs, 480 atoms) and the [C 2 mim] + cation with the labeled H typessolution, the ensemble-averaged 1 H chemical shifts (over a few MD snapshots) obtained with the PBC-GEBF approach at the B97-2/ pcSseg-2 level are in good agreement with the experimental results 66. Hence, to simulate the 1 H chemical shifts of [C 2 mim][BF 4 ] ILs, some snapshots of this ILs were also equidistantly selected along the MD trajectories.…”

Building Machine Learning Force Fields of Proteins with Fragment-Based Approach and Data Transfer

“…Due to the chemical complexities of proteins and the high computational costs of QM methods for large systems, building MLFFs for proteins remains a great challenge. Energy-based fragmentation (EBF) approaches [35][36][37][38][39][40][41][42][43][44][45] provide a practical and attractive solution to overcome these two difficulties. With this approach, the ground-state MLFF of a large system can be obtained as the linear combination of MLFF trained from small subsystems, which are representations of different local regions of a large system.…”

Building Machine Learning Force Fields of Proteins with Fragment-Based Approach and Transfer Learning

“…Due to the chemical complexities of proteins and high computational costs of QM methods for large systems, building MLFFs for proteins remains a great challenge. Energy-based fragmentation (EBF) approaches [28][29][30][31][32][33][34][35][36][37][38] provide a practical and attractive solution to overcome these two difficulties. With this approach, the ground-state MLFF of a large system can be obtained as the linear combination of MLFF trained from small subsystems, which are representation of different local regions of a large system.…”