Journal of Chemical Theory and Computation最新文献

{"title":"Benchmarking and Contrasting Exchange-Correlation Functional Differences in Response to Static Correlation in Unrestricted Kohn-Sham and a Hybrid 1-Electron Reduced Density Matrix Functional Theory.","authors":"Daniel Gibney, Jan-Niklas Boyn","doi":"10.1021/acs.jctc.5c00244","DOIUrl":"10.1021/acs.jctc.5c00244","url":null,"abstract":"<p><p>A hybrid Kohn-Sham Density Functional Theory (KS-DFT) and 1-electron Reduced Density Matrix Functional Theory (1-RDMFT) has recently been developed to describe strongly correlated systems at mean-field computational cost. This approach relies on combining a Reduced Density Matrix Functional to capture strong correlation effects with existing exchange correlation (XC) functionals to capture the remaining dynamical correlation effects. In this work, we systematically benchmark the performance of nearly 200 different XC functionals available within LibXC in this DFA 1-RDMFT framework, contrasting it with their performance in unrestricted KS-DFT. We identify optimal XC functionals for use within DFA 1-RDMFT and elucidate fundamental trends in the response of different XC functionals to strong correlation in both DFA 1-RDMFT and UKS-DFT.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":" ","pages":"5132-5142"},"PeriodicalIF":5.7,"publicationDate":"2025-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143952874","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Discrete Generalized Quantum Master Equations.","authors":"Nancy Makri","doi":"10.1021/acs.jctc.5c00396","DOIUrl":"10.1021/acs.jctc.5c00396","url":null,"abstract":"<p><p>Several derivative and integral approximations are explored for discretizing the Nakajima-Zwanzig generalized quantum master equation (NZ-QME or GQME) to obtain discrete quantum master equation (DQME) hierarchies and relations between discrete memory kernel and reduced density matrix (RDM) elements. It is shown that the simplest forward-difference approximation does not allow the reliable determination of the discrete kernel elements, even in the infinitesimal time-step limit, and that discrete kernels obtained in earlier work are flawed, although the procedure can be remedied. The various approximations give rise to DQMEs that differ in structure and in the RDM-kernel relationships. It is shown that the use of a more accurate discretization based on the midpoint derivative and midpoint integral approximations leads to a DQME that exhibits endpoint effects, which reflect the weaker impact of the bath on the RDM during the first time step and which parallel those encountered in the small matrix decomposition of the path integral (SMatPI) with a symmetric factorization of the short-time propagator. The features of the DQME hierarchies and RDM-kernel relations are illustrated through analytical examples involving a simple integrodifferential equation and a scalar GQME model, as well as numerical results for a two-level system (TLS) coupled to a harmonic bath.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":" ","pages":"5037-5048"},"PeriodicalIF":5.7,"publicationDate":"2025-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143956777","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multiobjective Evolutionary Strategy for Improving Semiempirical Hamiltonians in the Study of Enzymatic Reactions at the QM/MM Level of Theory.","authors":"José Luís Velázquez-Libera, Rodrigo Recabarren, Esteban Vöhringer-Martinez, Yamisleydi Salgueiro, J Javier Ruiz-Pernía, Julio Caballero, Iñaki Tuñón","doi":"10.1021/acs.jctc.5c00247","DOIUrl":"10.1021/acs.jctc.5c00247","url":null,"abstract":"<p><p>Quantum mechanics/molecular mechanics (QM/MM) simulations are crucial for understanding enzymatic reactions, but their accuracy depends heavily on the quantum-mechanical method used. Semiempirical methods offer computational efficiency but often struggle with accuracy in complex systems. This work presents a novel multiobjective evolutionary strategy for optimizing semiempirical Hamiltonians, specifically designed to enhance their performance in enzymatic QM/MM simulations while remaining broadly applicable to condensed-phase systems. Our methodology combines automated parameter optimization, targeting ab initio or density functional theory (DFT)-reference potential energy surfaces, atomic charges, and gradients, with comprehensive validation through minimum free energy path (MFEP) calculations. To demonstrate its effectiveness, we applied our approach to improve the GFN2-xTB Hamiltonian using two enzymatic systems that involve hydride transfer reactions where the activation energy barrier is severely underestimated: Crotonyl-CoA carboxylase/reductase (CCR) and dihydrofolate reductase (DHFR). The optimized parameters showed significant improvements in reproducing potential and free energy surfaces, closely matching higher-level DFT calculations. Through an efficient two-stage optimization process, we first developed parameters for CCR using reaction path data, then refined these parameters for DHFR by incorporating a targeted set of additional training geometries. This strategic approach minimized the computational cost while achieving accurate descriptions of both systems, as validated through QM/MM simulations using the Adaptive String Method (ASM). Our method represents an efficient approach for optimizing semiempirical methods to study larger systems and longer time scales, with potential applications in enzymatic reaction mechanism studies, drug design, and enzyme engineering.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":" ","pages":"5118-5131"},"PeriodicalIF":5.7,"publicationDate":"2025-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144042811","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Data-Driven Kinetic Reaction Networks for Separation Chemistry.","authors":"Jiyoung Lee, Logan J Augustine, Graeme Henkelman, Ping Yang, Danny Perez","doi":"10.1021/acs.jctc.4c01783","DOIUrl":"10.1021/acs.jctc.4c01783","url":null,"abstract":"<p><p>Understanding complex, multistep chemical reactions at the molecular level is a major challenge whose solution would greatly benefit the design and optimization of numerous chemical processes. The separation of rare-earth (4f) and actinide (5f) elements is an example where improving our chemical understanding is important for designing and optimizing new chemistries, even with a limited number of observations. In this work, we leverage data-driven artificial intelligence and machine-learning approaches to develop kinetic reaction networks that describe the liquid-liquid extraction mechanism of uranium using <i>N</i>,<i>N</i>-di-2-ethylhexyl-isobutyramide (DEHiBA). Specifically, we compare and contrast the properties of two classes of models: (1) purely data-driven models that are regularized using chemistry-agnostic, L1 regression and (2) chemistry-informed models that are regularized using relative reaction energies provided by quantum mechanical calculations. We observe that purely data-driven models are unbiased, simple, and accurate in their predictions of experimental measurements when provided with sufficient data but are difficult to fully constrain and interpret. In contrast, chemistry-informed models exhibit significantly improved chemical interpretability and consistency, providing a detailed description of the separation process while achieving high accuracy through ensemble averaging. Overall, the dominant species predicted to be extracted into the organic phase is UO₂(NO₃)₂(DEHiBA)₂, agreeing with experimental slope analysis, thermodynamic modeling, EXAFS, and crystal structures. This work demonstrates that leveraging the fundamental structure of the problem can lead to efficient learning schemes that provide both accurate predictions and chemical insights at a low computational cost.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":" ","pages":"5182-5193"},"PeriodicalIF":5.7,"publicationDate":"2025-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143955902","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Hybrid Transformer Architecture with a Quantized Self-Attention Mechanism Applied to Molecular Generation.","authors":"Anthony M Smaldone, Yu Shee, Gregory W Kyro, Marwa H Farag, Zohim Chandani, Elica Kyoseva, Victor S Batista","doi":"10.1021/acs.jctc.5c00331","DOIUrl":"10.1021/acs.jctc.5c00331","url":null,"abstract":"<p><p>The success of the self-attention mechanism in classical machine learning models has inspired the development of quantum analogs aimed at reducing the computational overhead. Self-attention integrates learnable <i>query</i> and <i>key</i> matrices to calculate attention scores between all pairs of tokens in a sequence. These scores are then multiplied by a learnable <i>value</i> matrix to obtain the output self-attention matrix, enabling the model to effectively capture long-range dependencies within the input sequence. Here, we propose a hybrid quantum-classical self-attention mechanism as part of a transformer decoder, the architecture underlying large language models (LLMs). To demonstrate its utility in chemistry, we train this model on the QM9 dataset for conditional generation, using SMILES strings as input, each labeled with a set of physicochemical properties that serve as conditions during inference. Our theoretical analysis shows that the time complexity of the query-key dot product is reduced from <math><mi>O</mi><mrow><mo>(</mo><msup><mrow><mi>n</mi></mrow><mrow><mn>2</mn></mrow></msup><mi>d</mi><mo>)</mo></mrow></math> in a classical model to <math><mi>O</mi><mrow><mo>(</mo><msup><mrow><mi>n</mi></mrow><mrow><mn>2</mn></mrow></msup><mi>log</mi><mo>⁡</mo><mi>d</mi><mo>)</mo></mrow></math> in our quantum model, where <i>n</i> and <i>d</i> represent the sequence length and the embedding dimension, respectively. We perform simulations using NVIDIA's CUDA-Q platform, which is designed for efficient GPU scalability. This work provides a promising avenue for quantum-enhanced natural language processing (NLP).</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":" ","pages":"5143-5154"},"PeriodicalIF":5.7,"publicationDate":"2025-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143955261","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Kinetics and Mechanism of Phase Separation in Ternary Lipid Mixtures Containing APP C99: Atomistic vs Coarse-Grained MD Simulations.","authors":"George A Pantelopulos, Sangram Prusty, Asanga Bandara, John E Straub","doi":"10.1021/acs.jctc.5c00016","DOIUrl":"10.1021/acs.jctc.5c00016","url":null,"abstract":"<p><p>The phase separation of lipid bilayers, composed of mixtures of saturated and unsaturated lipids and cholesterol, is a topic of fundamental importance in membrane biophysics and cell biology. The formation of lipid domains, including liquid-disordered domains enriched in unsaturated lipids and liquid-ordered domains enriched in saturated lipids and cholesterol, is believed to be essential to the function of many membrane proteins. Experiment, theory, and simulation have been used to develop a general understanding of the thermodynamic driving forces underlying phase separation in ternary and quaternary lipid mixtures. However, the kinetics of early events in lipid phase separation in the presence of transmembrane proteins remain relatively understudied. Using large-scale all-atom and coarse-grained simulations, we explore the kinetics and phase separation of ternary lipid mixtures of saturated lipid, unsaturated lipid, and cholesterol in the presence of transmembrane proteins. Order parameters employed in the Cahn-Hilliard theory provide insight into the kinetics and mechanism of lipid phase separation. We observe three distinct time regimes in the phase separation process: a shorter exponential time phase, followed by a power-law phase, and then a longer plateau phase. Comparison of lipid, protein, and lipid-protein dynamics between all-atom and coarse-grained models identifies both quantitative and qualitative differences and similarities in the phase separation kinetics. Moreover, timescaling of the dynamics of the AA and CG simulations yields a similar kinetic mechanism of phase separation. The findings of this study elucidate fundamental aspects of membrane biophysics and contribute to ongoing efforts to define the role of lipid rafts in the structure and function of the cellular membrane.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":" ","pages":"5291-5303"},"PeriodicalIF":5.7,"publicationDate":"2025-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144074924","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Computing Excited States of Molecules Using Normalizing Flows.","authors":"Yahya Saleh, Álvaro Fernández Corral, Emil Vogt, Armin Iske, Jochen Küpper, Andrey Yachmenev","doi":"10.1021/acs.jctc.5c00590","DOIUrl":"10.1021/acs.jctc.5c00590","url":null,"abstract":"<p><p>Calculations of highly excited and delocalized molecular vibrational states are computationally challenging tasks, which strongly depend on the choice of coordinates for describing vibrational motions. We introduce a new method that leverages normalizing flows, i.e, parametrized invertible functions, to learn optimal vibrational coordinates that satisfy the variational principle. This approach produces coordinates tailored to the vibrational problem at hand, significantly increasing the accuracy and enhancing the basis set convergence of the calculated energy spectrum. The efficiency of the method is demonstrated in calculations of the 100 lowest excited vibrational states of H₂S, H₂CO, and HCN/HNC. The method effectively captures the essential vibrational behavior of molecules by enhancing the separability of the Hamiltonian and hence allows for an effective assignment of approximate quantum numbers. We demonstrate that the optimized coordinates are transferable across different levels of basis set truncation, enabling a cost-efficient protocol for computing vibrational spectra of high-dimensional systems.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":" ","pages":"5221-5229"},"PeriodicalIF":5.7,"publicationDate":"2025-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12120919/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144074921","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"IPAMD: A Plugin-Based Software for Biomolecular Condensate Simulations","authors":"Xiao-Yang Liu,&nbsp;You-Liang Zhu,&nbsp;Yu-Ze Jiang,&nbsp;Shao-Kang Shi,&nbsp;Li Zhao* and Zhong-Yuan Lu*,&nbsp;","doi":"10.1021/acs.jctc.5c0014710.1021/acs.jctc.5c00147","DOIUrl":"https://doi.org/10.1021/acs.jctc.5c00147https://doi.org/10.1021/acs.jctc.5c00147","url":null,"abstract":"<p >The study of intrinsically disordered proteins (IDPs) and their role in biomolecular condensate formation has become a critical area of research, offering insights into fundamental biological processes and therapeutic development. Here, we present IPAMD (Intrinsically disordered Protein Aggregation Molecular Dynamics), a plugin-based software designed to simulate the formation dynamics of biomolecular condensates of IDPs. IPAMD provides a modular, efficient, and customizable simulation platform specifically designed for biomolecular condensate studies. It incorporates advanced force fields, such as HPS-based and Mpipi models, and employs optimization techniques for large-scale simulations. The software features a user-friendly interface and supports batch processing, making it accessible to researchers with varying computational expertise. Benchmarking and case studies demonstrate the ability of IPAMD to accurately simulate and analyze condensate structures and properties.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":"21 11","pages":"5746–5756 5746–5756"},"PeriodicalIF":5.7,"publicationDate":"2025-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144238726","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Advancing Multiscale Molecular Modeling with Machine Learning-Derived Electrostatics.","authors":"Jonathan A Semelak, Ignacio Pickering, Kate Huddleston, Justo Olmos, Juan Santiago Grassano, Camila Mara Clemente, Salvador I Drusin, Marcelo Marti, Mariano Camilo Gonzalez Lebrero, Adrian E Roitberg, Dario A Estrin","doi":"10.1021/acs.jctc.4c01792","DOIUrl":"10.1021/acs.jctc.4c01792","url":null,"abstract":"<p><p>We introduce an innovative machine learning (ML)-based framework for multiscale molecular modeling in which the ML subsystem is treated as an electrostatic entity interacting with its molecular mechanics (MM) environment through classical electrostatics. The integration of ML accuracy with multiscale modeling is accomplished by leveraging the capabilities of the ANI neural networks to predict geometry-dependent atomic partial charges at the minimal basis iterative stockholder (MBIS) level, going beyond static mechanical embedding. This ML/MM approach can closely approximate state-of-the-art multiscale quantum-classical (QM/MM) methods while significantly lowering computational requirements, thereby facilitating more efficient and precise simulations in computational chemistry. The method requires no additional training beyond the initial model setup and is integrated into Amber, one of the most widely used software suites for molecular modeling, ensuring accessibility to the broader community. We validate its performance across a variety of challenging applications, including the solvation structure, vibrational spectra, torsion free energy profiles, and protein-ligand interactions, achieving excellent agreement with QM/MM benchmarks. This framework not only advances the frontiers of multiscale modeling but also showcases the potential of machine learning to achieve quantum-level accuracy with exceptional efficiency for complex chemical systems.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":" ","pages":"5194-5207"},"PeriodicalIF":5.7,"publicationDate":"2025-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143555340","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multistep Approach for Simulating Raman Spectra of Amorphous Materials: The Case of Li3PS4 Glass Electrolyte","authors":"Jakub Pawelko*,&nbsp;Eric Furet,&nbsp;Gwenael Duplaix-Rata,&nbsp;Nicolas Perrin and Xavier Rocquefelte*,&nbsp;","doi":"10.1021/acs.jctc.5c0032110.1021/acs.jctc.5c00321","DOIUrl":"https://doi.org/10.1021/acs.jctc.5c00321https://doi.org/10.1021/acs.jctc.5c00321","url":null,"abstract":"<p >Glasses are widely used for their various applications, which arise from their inherent lack of long-range ordering. This characteristic makes it challenging to describe their atomic properties. To facilitate and accelerate glass research, computational simulations, such as molecular dynamics or Monte Carlo simulations, are commonly employed to model the structure of these amorphous materials. However, verifying and later refining the models require comparing these simulations with spectroscopic data, which can be quite computationally challenging due to the number of atoms needed in the unit cell to account for the disorder. In this paper, we propose a multistep approach for simulating specifically Raman spectra that accounts for long-range interactions in amorphous materials. This approach, going from molecular dynamics to embedded cluster calculations, allows us to define the Raman signature of each structural unit and reconstruct the Raman spectrum of the glass electrolyte Li₃PS₄, achieving low computational cost and high agreement with existing spectroscopic experimental data.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":"21 11","pages":"5679–5685 5679–5685"},"PeriodicalIF":5.7,"publicationDate":"2025-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144239128","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}