Journal of Chemical Theory and Computation最新文献

{"title":"Limitations of Fixed-Point-Charge Force Fields for Metal Coordination Modes in Proteins: A Multilevel Study of Ca2+ in Integrin αVβ3","authors":"Andrea Levy,&nbsp;Giulia Frigerio,&nbsp;Jules Grollier,&nbsp;Paulo Siani,&nbsp;Cristiana Di Valentin,&nbsp;Ursula Rothlisberger","doi":"10.1021/acs.jctc.6c01320","DOIUrl":"10.1021/acs.jctc.6c01320","url":null,"abstract":"<p>Standard biomolecularforce fields often present limitations inmodeling metal coordination modes. We combined classical and QM/MMmolecular dynamics to investigate Ca²⁺-mediated bindingof cRGD to integrin α_Vβ₃. We showthat different metal coordination modes imply variations in the partialcharges of the metal-coordinating atoms, which, if incorporated inthe force field, substantially alter the free energy landscape. Theseresults highlight the value of QM-based multilevel approaches to accuratelycapture metal coordination modes in biomolecular systems.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":"22 21","pages":"11077–11086"},"PeriodicalIF":5.8,"publicationDate":"2026-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149275080","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 Complete Basis Set Limit Estimates from a Minimal Auxiliary Basis","authors":"Nicolas Grimblat,&nbsp;Gabriel Klassen,&nbsp;Guido Falk von Rudorff","doi":"10.1021/acs.jctc.6c01004","DOIUrl":"10.1021/acs.jctc.6c01004","url":null,"abstract":"<p>Quantum chemistry calculations are often performed usingatom-centeredbasis sets, which are chosen to balance accuracy and cost. While theyare systematically improvable, the total energy converges slowly withbasis set size toward the complete basis set (CBS) limit. Common extrapolationmethods require several intermediate-quality calculations to affordan estimate of the CBS energy. We propose a minimal complementaryauxiliary basis set (CABS) baseline to estimate the CBS energy froma single quantum chemistry calculation in a minimal basis set viakernel ridge regression (KRR), which is more efficient than both directand Δ-machine learning and 40× faster than triple-ζsingle-point calculations. The spatial and alchemical gradients fromminimal basis and minimal CABS are more accurate than double-ζ(DZ) calculations. We show that KRR on standard molecular representationscan be improved by approximating atom-wise local kernels using Chebyshevpolynomials, which allows us to train KRR models 10×–50×times faster on moderate compute resources, thereby enabling a data-drivenapproach toward the mean-field CBS, combining physical baselines capturingleading-order effects with data-efficient machine learning models.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":"22 21","pages":"11277–11286"},"PeriodicalIF":5.8,"publicationDate":"2026-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149210879","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":"Investigation of Magnetic Properties in Exchange Coupled Radical-Nickelocene Complexes by Adopting a Generalized Spin Hamiltonian","authors":"Abhishek R. Nath,&nbsp;Md. Ehesan Ali","doi":"10.1021/acs.jctc.6c00939","DOIUrl":"10.1021/acs.jctc.6c00939","url":null,"abstract":"<p>Magnetic interaction between an isotropic spin centerand a stronglyanisotropic partner presents a unique challenge in molecular magnetism,as the interplay between local anisotropy and exchange pathways oftendefies conventional Heisenberg exchange. In this work, we addressthis complexity by investigating a series of nitronyl-nitroxide radicalscoupled to nickelocene through polyene spacers of increasing length.Both density functional theory and wave function based multireferenceapproaches are adopted to investigate the nature of isotropic exchangein these systems. To resolve the limitations of standard spin models,we employ a mapping approach based on an effective Hamiltonian derivedfrom CASSCF + NEVPT2 spin–orbit eigenstates. By treating thesystem’s low-lying electronic structure through this extendedframework, we explicitly account for the significant mixing betweenthe ground and first excited spin states which is found to be responsiblefor an anomalous increase in the axial zero-field splitting (D). Wefind that increasing the model space to include antiferromagneticstates recovers the full zero-field splitting while enabling finiteanisotropic exchange contributions to be extracted. By establishinga systematic distance dependence for these interactions, this workprovides a rigorous framework for navigating the nontrivial couplingmechanisms in complex organic–inorganic spin architectures.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":"22 21","pages":"11266–11276"},"PeriodicalIF":5.8,"publicationDate":"2026-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149241654","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":"SPectroscopy Informed Configurational Enhanced Sampling (SPICES)","authors":"Roel van de Ven,&nbsp;Herma Cuppen,&nbsp;Daria Ruth Galimberti","doi":"10.1021/acs.jctc.6c01326","DOIUrl":"10.1021/acs.jctc.6c01326","url":null,"abstract":"<p>We present a newenhanced-sampling scheme for obtaining the underlyingconformer starting from a spectrum: the SPectroscopy Informed ConfigurationalEnhanced Sampling (SPICES) method. SPICES is driven by a history-dependentbiased molecular dynamics, in which the biases are governed by spectralinformation rather than by energetic considerations. The performanceof SPICES is hence rather insensitive to the target conformer’srelative energy compared to the global minimum structure. Furthermore,it does not require pretraining or prior knowledge of the system.We have critically assessed the performance of SPICES and demonstratedthat it can identify the correct conformer of our test system, diphenylalanine(FF), among the 400 stable conformations in 22 spectral calculations(median value). We also show that SPICES is robust to noise in theexperimental spectrum and to small discrepancies between theory andexperiment.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":"22 21","pages":"11471–11483"},"PeriodicalIF":5.8,"publicationDate":"2026-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149241768","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":"Quantitative Pathway-Resolved Kinetics from Neural Network-Guided Weighted Ensemble Simulations","authors":"Dibyendu Maity,&nbsp;Shaheerah Shahid,&nbsp;Sayari Bhattacharya,&nbsp;Rupak Majumdar,&nbsp;Suman Chakrabarty","doi":"10.1021/acs.jctc.6c01321","DOIUrl":"10.1021/acs.jctc.6c01321","url":null,"abstract":"<p>Many crucial biomolecular processes, such as ligand dissociationand conformational switching, occur via multiple competing transitionpathways. While standard rare-event sampling methodologies providerigorous estimates of the global transition rate, they typically collapsethe underlying mechanistic heterogeneity into a single global constant,obscuring the relative kinetic contributions of distinct channels.Here, we introduce a path-resolved weighted ensemble (WE) frameworkthat quantitatively assigns steady-state flux to competing transitionchannels without modifying the underlying unbiased dynamics. In ourapproach, initial reactive trajectories generated via direction-guidedadaptive sampling are clustered by dynamic time warping (DTW) andsmoothed into neural-network-refined representative paths. These refinedroutes are formulated as path collective variables (PathCVs) to preseedand guide channel-specific WE simulations. We demonstrate the robustnessof this framework across systems of increasing complexity. On a modelthree-hole potential, the method accurately captures a temperature-dependentpathway crossover. In the benzamidine–trypsin complex, it isolatesthe dominant unbinding route while quantifying the fluxes of subdominantchannels. Finally, for imatinib dissociation from Abl kinase, themethod reveals that the clinically relevant N368S resistance mutationdrastically redistributes reactive flux from the P-loop/hinge channeltoward the αC-helix channel, accelerating escape. This frameworksuccessfully translates qualitative pathway classification into rigorous,pathway-specific kinetic observables, offering a powerful tool forunderstanding mutation-induced mechanistic shifts in biomolecularfunction.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":"22 21","pages":"11111–11124"},"PeriodicalIF":5.8,"publicationDate":"2026-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149251037","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":"Global Potential Energy Surface Exploration Using Generative Stochastic Surface Walking (Gen-SSW)","authors":"Qiming Liang,&nbsp;Zi-Xing Guo,&nbsp;Jin-Peng Tang,&nbsp;Tong Guan,&nbsp;Cheng Tong,&nbsp;Ye-Fei Li,&nbsp;Zhi-Pan Liu","doi":"10.1021/acs.jctc.6c01420","DOIUrl":"10.1021/acs.jctc.6c01420","url":null,"abstract":"<p>Globalpotential energy surface (PES) exploration offers a uniqueway to predict the structure of material from first principles. Here,we develop Gen-SSW method to enable autonomous and fast global PESexploration, which combines a general-purpose SE(3)-equivariant 3Ddiffusion model, LASP-Gen, with stochastic surface walking (SSW) globaloptimization. Contrary to the conventional SSW global PES explorationthat requires a long serial trajectory to overcome high barriers betweenlarge structural basins, Gen-SSW establishes an autonomous, massivelyparallel global optimization framework, which realizes global PESexploration via divide-and-conquer: many short SSW trajectories areperformed based on the seed structures batch-generated from LASP-Gen.By applying Gen-SSW across complex materials, we show that (i) LASP-Gencan efficiently yield a structural pool with huge structure diversityand reasonable local chemical bonding patterns, and (ii) By dockingLASP-Gen with SSW, Gen-SSW can identify efficiently low-energy polymorphsof materials, as exampled by ionic materials (TiO₂, LiFePO₄), amorphous alloys (Ge₂Sb₂Te₅), strong covalent materials (boron) and molecular crystals (phenolcrystal). In particular, we find a new boron allotropy, namely α̃-B,that is only 6 meV/atom above α-boron but 17 meV/atom more stablethan γ-boron, and suggest a synthetical route for making thisnew phase via solid–solid phase transition.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":"22 21","pages":"11705–11718"},"PeriodicalIF":5.8,"publicationDate":"2026-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149304204","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":"Multiplicative Error Analysis: A Scale-Invariant View of TDDFT Two-Photon Absorption Benchmarks","authors":"Dávid Mester,&nbsp;Mihály Kállay","doi":"10.1021/acs.jctc.6c01476","DOIUrl":"10.1021/acs.jctc.6c01476","url":null,"abstract":"<p>Two-photon absorption (2PA) strengths span several ordersof magnitude,yet time-dependent density functional theory (TDDFT) benchmark studieshave commonly relied on raw-scale mean absolute errors and linearregression statistics. Such statistics can be dominated by the mostintense transitions and do not distinguish systematic multiplicativebias, typical fold error, dynamic range reproduction, and rank preservation.Here, published TDDFT 2PA benchmarks based on coupled-cluster referencesare reanalyzed by using a scale-invariant, multiplicative framework.For the most widely used benchmark set [<em>J. Chem.</em><em>Theory Comput.</em> 2022, 18, 1046], it is shown that the fivestrongest transitions, representing only approximately 10% of thedata, account for up to 65.5% of the total raw absolute error. Consequently,raw-scale metrics assign disproportionate weight to a small numberof intense transitions and may obscure method performance across theremainder of the intensity range. Across several benchmark sets, theanalysis further reveals that similar raw-scale errors can arise fromqualitatively different error structures. The multiplicative analysisgoes beyond previously recognized qualitative trends by resolvingthem into directly comparable and interpretable components. Theseresults demonstrate that neither individual raw-scale metrics northeir combinations provide a sufficient basis for identifying a reliableor generally recommended functional, and they support a multidimensional,scale-invariant assessment of 2PA benchmarks. The same framework mayalso be useful for other properties for which multiplicative errorsprovide a more appropriate description than additive deviations suchas oscillator strengths, kinetic and radiative rate constants, andequilibrium constants.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":"22 21","pages":"11378–11392"},"PeriodicalIF":5.8,"publicationDate":"2026-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149209799","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":"Performance of Neuroevolution Potentials for the Ionic Liquid [Emim][Ac]: A Systematic Benchmark of Four Density Functionals","authors":"Junfeng Lu,&nbsp;Yucong Ge,&nbsp;Yanlei Wang,&nbsp;Hongyan He","doi":"10.1021/acs.jctc.6c01816","DOIUrl":"10.1021/acs.jctc.6c01816","url":null,"abstract":"<p>Machine-learned force fields (MLFFs)based on NeuroevolutionPotential(NEP) models were developed for the ionic liquid 1-ethyl-3-methylimidazoliumacetate ([Emim][Ac]) using first-principles reference data generatedwith four density functionals: PBE-D3, RPBE-D3, optB88-vdW, and SCAN+rVV10.Using a unified training protocol, the resulting NEP models were systematicallybenchmarked against structural, thermophysical, transport, and hydrogen-bond-networkproperties, including radial distribution functions (RDFs), coordinationnumbers, neutron total scattering profiles, density, viscosity, ionicconductivity, and thermal conductivity. Spearman correlation analysisand weighted composite scoring were further used to establish a property-orientedevaluation framework for MLFF assessment and selection. The resultsreveal a pronounced dependence of NEP performance on the density functionalused to generate the reference data. Although all four models reproducethe dominant local structural features of [Emim][Ac], appreciabledifferences emerge in medium-range structural ordering and transportproperties. The PBE-D3-based NEP provides the most accurate overalldescription of structural properties, whereas the optB88-vdW-basedNEP yields the closest agreement with experimental transport properties,particularly viscosity and ionic conductivity. Analysis of the hydrogen-bondnetwork further reveals systematic relationships among hydrogen-bondpopulations, residence times, free-ion fractions, and ion transport.These results demonstrate that, for [Emim][Ac], the exchange–correlationfunctional used for reference-data generation influences the physicalfidelity and property-specific performance of the resulting NEP models.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":"22 21","pages":"11595–11607"},"PeriodicalIF":5.8,"publicationDate":"2026-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149226499","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":"Predicting Ternary Complex Stability through Binding Free-Energy Calculations in the Geometric Route","authors":"Benedict Tan,&nbsp;Alessio Ciulli,&nbsp;Julien Michel","doi":"10.1021/acs.jctc.6c01541","DOIUrl":"10.1021/acs.jctc.6c01541","url":null,"abstract":"<p>Molecular glue-induced targeted protein degradation isa powerfulemerging therapeutic modality that exploits the ubiquitin-proteasomedegradation pathway to break down proteins associated with diseaseproliferation. At the structural level, degradation proceeds throughthe formation of a ternary complex induced by small molecules betweenthe target protein and an E3 ligase. Unlike PROTACs, molecular gluesdo not require a chemical linker and tend to stabilize preexistingnative protein–protein interactions. Computational screeningof novel glues is challenging, in large part due to the complexityand high degree of cooperativity involved in the formation of theternary complex interface. Here, we present a rigorous Molecular Dynamics-basedsimulation protocol that uses Umbrella Sampling to calculate the ternarycomplex binding free energy. The protocol follows a geometric thermodynamiccycle with conformational and Boresch-style orientational restraintsto drive the system from bound to unbound. Our framework successfullypredicts the effect of single-point interface mutations on the relativestability of the CK1α-lenalidomide-CRBN ternary complex (<em>R</em>² = 0.95 compared to experimental binding affinities).We also study a series of systems related to the BRD4 degraders IBG1and IBG3, which recruit the DCAF16 E3 ligase. We rank various complexesengaging DCAF16 by their binding free energies and achieve an orderingthat is consistent with the trends observed in biophysical assays.Of particular note is our ability to assert the reliance of BRD4 selectivityon a single interfacial glycine residue, which alternative, reduced-samplingMM/GBSA or HADDOCK scoring methodologies fail to achieve. Overall,our protocol shows promise as a generalizable approach to rankingternary complexes by thermodynamic stability, potentially aiding inthe identification of potent new degraders.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":"22 21","pages":"11662–11675"},"PeriodicalIF":5.8,"publicationDate":"2026-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149229969","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":"Beyond Multiscale: A Neural-Network-Unified Molecular Model for Asymmetry-Free Dynamics Simulation","authors":"Takuma Ikeda,&nbsp;Hiroshi C. Watanabe,&nbsp;Haruyuki Nakano","doi":"10.1021/acs.jctc.6c01037","DOIUrl":"10.1021/acs.jctc.6c01037","url":null,"abstract":"<p>Adaptive multiscalesimulations, such as adaptive quantum mechanics(QM)/molecular mechanics (MM), are indispensable for investigatingcomplex solution dynamics but historically suffer from a fundamentaldilemma: the dynamic exchange of identical solvent molecules acrossdifferent model resolutions introduces an energetic asymmetry undersolvent exchange (specifically, a permutation asymmetry where physicallyidentical molecules are treated indistinguishably yet subjected todifferent force fields depending on their instantaneous positions).This asymmetry forces a prohibitive trade-off between severe structuraldistortions and the violation of Hamiltonian conservation (continuousenergy drift). To resolve this issue, we propose the “IntegratedMolecular Model,” a paradigm-shifting framework that unifiesmultiple resolutions into a single, dynamically consistent potential.By employing a neural network that takes only the multiscale weightfunctions as input descriptors, our model learns to systematicallycancel the nonphysical transition forces originating from spatialdiscontinuities. As a proof-of-concept, we validated this unifiedframework on bulk water, aqueous ionic solutions, and reactive hydroniumions. Our approach not only faithfully restores artifact-free solvationstructures but also strictly conserves the Hamiltonian, enabling stable,long-term simulations under the microcanonical (NVE) ensemble. Consequently,the model successfully eliminates artificial thermostat-induced breathingmodes, allowing for the rigorous evaluation of intrinsic dynamicalproperties such as diffusion coefficients. Furthermore, owing to itsphysically informed network design, the correction potential introducesless than 2% additional computational overhead and demonstrates remarkablezero-shot transferability across chemically similar ionic species.By eradicating the inherent asymmetry of adaptive multiscale modeling,this highly scalable framework establishes a robust foundation forhigh-fidelity simulations of complex condensed-phase phenomena, includingreactive dynamics at solid–liquid interfaces and ion transportin battery electrolytes.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":"22 21","pages":"11087–11100"},"PeriodicalIF":5.8,"publicationDate":"2026-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149091988","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}