Journal of Chemical Theory and Computation最新文献

{"title":"Localized and Delocalized Nonadiabatic Vibronic Models Unravel Chiroptical Trends and Design Principles in Symmetric π Conjugated Macrocycles","authors":"Francisco García-González,&nbsp;Julia Portela-Pino,&nbsp;Daniel Aranda,&nbsp;José Lorenzo Alonso-Gómez,&nbsp;Fabrizio Santoro","doi":"10.1021/acs.jctc.6c01519","DOIUrl":"10.1021/acs.jctc.6c01519","url":null,"abstract":"<p>The chiroptical response of π-conjugated macrocyclesassembledfrom repeating chromophores can be remarkably intense, yet how itscales with ring size, and what controls it, remains only partly understood.Here we investigate a series of axially chiral diethynylspirobifluorenemacrocycles, (<em>P</em>_<em>n</em>)-<strong>DES</strong>_<em>n</em> (<em>n</em> = 2,3, 4; <em>D</em>₂, <em>D</em>₃ and <em>D</em>₄ symmetry), combining accuratenonadiabatic quantum dynamics with two complementary linear vibroniccoupling Hamiltonians: a delocalized model, whose diabatic statesspan the whole macrocycle, and a localized, exciton-like model builtfrom the excitations of the individual fragments. Both reproduce theexperimental one-photon absorption and electronic circular dichroism(ECD) spectra and reveal that nonadiabatic interactions, conical intersectionsand (pseudo-)Jahn–Teller couplings govern their shapes; a hierarchicalrepresentation of the Hamiltonian keeps the simulations tractable,even for (<em>P</em>₄)-<strong>DES</strong>₄, with more than 500 vibrational modes. A simple symmetry analysisbridges the two pictures, tracing every coupling back to the singleexcited-state gradient of the monomer and partitions the magnetictransition dipole into an intrinsic contribution, from the strainedmonomer and an extrinsic, excitonic one. This decomposition rationalizesthe observed trends explaining why the ECD grows super-linearly withring size. As a result, the response crosses over from strain-dominatedchirality in (<em>P</em>₂)-<strong>DES</strong>_{<strong>2</strong>} to excitonic chirality in (<em>P</em>₄)-<strong>DES</strong>₄ at a nearly constant dissymmetryfactor. Crucially, the excitonic rotatory strengths form a nearlyconservative couplet whose oppositely signed branches partly cancel,and their splitting (0.12–0.16 eV) is comparable to the widthof the vibronic progressions. The intrinsic band shape of the monomer,not the splitting alone, therefore sets how much ECD survives. Wetranslate these findings into geometric, excitonic and vibronic designprinciples for <em>g</em>-factor improved symmetric chiralmacrocycles, suggesting that our proposed computational approach canbe a fast screening tool for new designed macrocycles before synthesis.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":"22 21","pages":"11484–11497"},"PeriodicalIF":5.8,"publicationDate":"2026-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149210880","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":"Differentiable OPLS Force-Field Parameterization for Ionic Electrolytes and High-Throughput Application to Lithium-Ion Batteries","authors":"Haichao Huang,&nbsp;Zilin Chen,&nbsp;Qi Liu,&nbsp;Tianqi Zhao,&nbsp;Yunpei Liu,&nbsp;Guotao Qiu,&nbsp;Jianhui Chen,&nbsp;Zhen Li,&nbsp;Wenshuo Liang,&nbsp;Minsung Cho,&nbsp;Manxue Zhang,&nbsp;Feiyu Kang,&nbsp;Xiaolong Zou,&nbsp;Yidan Cao,&nbsp;Xushan Zhao,&nbsp;Ziqi Cheng,&nbsp;Ye Mei","doi":"10.1021/acs.jctc.6c01451","DOIUrl":"10.1021/acs.jctc.6c01451","url":null,"abstract":"<p>The rational design of lithium-ion battery (LIB) electrolytesisseverely impeded by the vast chemical space of solvent–saltcombinations and the inefficiency of empirical trial-and-error approaches.Molecular dynamics (MD) bridges microscopic solvation structures andmacroscopic physicochemical properties, but classical force fields(FFs) often lack accuracy for multicomponent electrolytes. To addressthis, we develop an automated differentiable OPLS-AA FF parametrizationworkflow tailored for ionic electrolytes. It employs topology-guidedatom typing to reduce parameter redundancy and optimizes Lennard-Jonesparameters via the DMFF framework. A uniform charge-scaling schemeapproximates condensed-phase charge transfer and polarization. Densityserves as the structural/thermodynamic fitting target, while ionicconductivity provides the independent transport benchmark, ensuringbalanced accuracy. Rigorous convergence tests establish a standardizedprotocol with ∼100,000-atom systems and 35–40 ns NVTtrajectories for reliable transport property calculations. Furthermore,high-throughput simulations of over 10,000 formulations─encompassing67 solvents and 15 lithium salts─are performed on Tianqiong,a special-purpose supercomputer for molecular dynamics simulationsthat has been validated to deliver numerically equivalent resultsto GPUs with superior efficiency. A comprehensive data set is generatedcovering five key electrolyte properties. The t-SNE visualizationreveals partial clustering of distinct salt chemistries, continuousproperty gradients as functions of salt concentration and temperature,and internal physical self-consistency across the measured electrolyteparameters while simultaneously identifying solvent components asanother critical performance determinant. The accurate, transferableFF and the rich data resource establish a robust foundation for futuredata-driven design of ionic electrolytes.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":"22 21","pages":"11569–11582"},"PeriodicalIF":5.8,"publicationDate":"2026-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149221786","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":"Uncertainty Quantification for Free Energy Calculations by Generalized Hierarchical Bayesian Inference","authors":"Martin Skorna,&nbsp;Adam Gottfried,&nbsp;Zuzana Janackova,&nbsp;Katarina Baxova,&nbsp;Pavel Jungwirth,&nbsp;Brennon L. Shanks","doi":"10.1021/acs.jctc.6c01533","DOIUrl":"10.1021/acs.jctc.6c01533","url":null,"abstract":"<p>Free energy calculationsare routinely used to studymolecularprocesses inaccessible to unbiased molecular dynamics, but their utilityultimately depends on knowing when and how much their predictionscan be trusted. Uncertainty estimation is therefore essential fordistinguishing genuine physical features of a free energy profilefrom artifacts arising from limited simulation data or inadequatesampling. Gaussian processes have emerged as a powerful frameworkfor reconstructing free energy profiles together with predictive uncertainties.However, existing implementations typically condition on fixed hyperparametersand observation noise, preventing predictive uncertainties from adaptingto the information content of the simulation data. Here, we developa generalized hierarchical Gaussian process framework that accountsfor these neglected sources of uncertainty. Applications to umbrellasampling and extended Lagrangian metadynamics of peptide-lipid membraneinteractions demonstrate that the resulting uncertainty estimatestrack reconstruction errors across a wide range of sampling and dataconditions.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":"22 21","pages":"11221–11232"},"PeriodicalIF":5.8,"publicationDate":"2026-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149226483","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":"Δ-HGPR: A Hierarchical Gaussian Process Framework for Active-Learning Δ-Potentials","authors":"Takahito Nakajima","doi":"10.1021/acs.jctc.6c01498","DOIUrl":"10.1021/acs.jctc.6c01498","url":null,"abstract":"<p>Δ-learning,which uses machine learning to represent onlythe difference between a high-accuracy reference and a lower-accuracybase calculator, is a promising route to high-fidelity molecular dynamics(MD) with limited training data, but it has lacked a unified protocolcombining hierarchical descriptors with well-defined equilibrium sampling.We propose Δ-HGPR (hierarchical Gaussian process regressionfor Δ-learning), a framework for learning difference potentials,and its MD instantiation, Δ-HGPR-MD. The framework combines(i) a difference potential written as an additive radial (inverse-distance)term plus a many-body machine-learning correction, determined by asingle regularized least-squares fit; (ii) the ensemble force standarddeviation as an active-learning indicator with an automatically scaledthreshold; and (iii) a difference potential held fixed within eachround so that equilibrium NVT sampling is well-defined in both explorationand production. The many-body term of Δ-HGPR-MD uses smoothoverlap of atomic positions (SOAP) descriptors with random Fourierfeatures (Δ-HGPR-RFF). On a toy system whose difference potentialcontains a 1/<em>r</em> contribution beyond the cutoff ofthe local descriptor, the many-body term alone reproduces none ofthis force, however many features it is given, whereas the hierarchyrecovers 99% of it, with a reach set by the cutoff of its radial term.In MD, the hierarchy reproduces about 80% of the shift of the ion-pairdistribution toward the reference, while the many-body term aloneleaves it where the base places it. On malonaldehyde and the Zundelcation <span><span></span><math><msub><mrow><mi>H</mi></mrow><mrow><mn>5</mn></mrow></msub><msubsup><mrow><mi>O</mi></mrow><mrow><mn>2</mn></mrow><mrow><mo>+</mo></mrow></msubsup></math></span> with three base calculators, Δ-HGPR-RFFremoves barrier errors of up to 0.1 eV under one protocol, givingfree-energy barriers statistically indistinguishable from DFT; free-energyperturbation shows that the error contributed by the learned correctionis below 0.01 eV on average in every case. For <span><span></span><math><msub><mrow><mi>H</mi></mrow><mrow><mn>5</mn></mrow></msub><msubsup><mrow><mi>O</mi></mrow><mrow><mn>2</mn></mrow><mrow><mo>+</mo></mrow></msubsup></math></span>with ANI-2x, whose dynamicscollapse withoutcorrection, it recovers the Zundel geometry and a barrier and transitioncount consistent with DFT.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":"22 21","pages":"11153–11167"},"PeriodicalIF":5.8,"publicationDate":"2026-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149241653","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":"Chemical Similarity Enables a Unified Free-Energy Framework for f-Block Solvation Thermodynamics","authors":"Hossam Farag,&nbsp;Michael J. Servis","doi":"10.1021/acs.jctc.6c01464","DOIUrl":"10.1021/acs.jctc.6c01464","url":null,"abstract":"<p><em>f</em>-Block elements underpin many essentialtechnologies,yet their efficient mutual separation remains a foundational challengein chemistry. The lanthanide series provide a stringent test case:the trivalent lanthanide ions (Ln³⁺) exhibit near-identicalsolvation chemistry across the series, distinguished only by the lanthanidecontraction. Here, we turn this chemical similarity into a computationaladvantage. Treating each solvated Ln³⁺ as a distinct alchemicalstate, we exploit configurational overlap between neighboring ionsby applying the multistate Bennett acceptance ratio (MBAR) to estimaterelative hydration free energies from pooled ensembles. We extendthis framework to biased simulations by coupling coordination-number-basedwell-tempered metadynamics with MBAR, enabling enhanced sampling ofrare coordination fluctuations within a unified free-energy estimator.Benchmarked in water as a controlled, well-characterized solvent,the workflow yields convergence, overlap, and effective sample size(ESS) diagnostics alongside free-energy trends in good agreement withexperiment. MBAR-enabled energy–entropy decomposition revealsdiscrepancies in the underlying energetic and entropic contributionsdespite good agreement in total free energy, motivating a multifidelitytest with a machine-learned polarizable interatomic potential. MBAR-basedreweighting reveals insufficient Hamiltonian overlap, exposing a transferabilitybottleneck; the pooled classical ensemble does not support reliablepost hoc reweighting to the target Hamiltonians. Domain-specific fine-tuning,renewed sampling, and explicit overlap reassessment are thereforerequired before MACE-polar can be incorporated as a higher-fidelitylayer. This framework establishes “U<em>f</em>S”─aunified free-energy framework for <em>f</em>-block solvation─asa scalable, uncertainty-aware approach to lanthanide solvation thermodynamicsand a foundation for future applications to related actinides, speciation-richenvironments, and data-driven design of new separations chemistries.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":"22 21","pages":"11206–11220"},"PeriodicalIF":5.8,"publicationDate":"2026-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149251049","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":"Periodic Electron Density Overlap Descriptors for Crystal Band Gap Prediction","authors":"Zhaosheng Zhang,&nbsp;Yiheng Xi,&nbsp;Jiadong Liu,&nbsp;Yanbo Liu","doi":"10.1021/acs.jctc.6c02070","DOIUrl":"10.1021/acs.jctc.6c02070","url":null,"abstract":"<p>A valence electron density encodes spatial electronicstructurethat complements an atomic position representation. We examine whethera periodic adaptation of the smooth overlap of electron densities(SOED) improves the modeling of density functional theory band gapsin inorganic crystals. Atom-centered neighborhoods are sampled directlyfrom periodic valence density grids, expanded in radial functionsand spherical harmonics, and summarized by rotationally invariantpower spectra. A scaffold density constructed from atomic positionssupplies a SOAP-inspired control with the same overlap machinery.We evaluate 40490 structures from MP-20-Charge after filtering undefinedPauling electronegativities and repartitioning by reduced formula.A validation-selected electronic SOED ensemble attains a held-outRMSE of 0.6179 eV, compared with 0.6595 eV for a local chemistry-matchedSOAP control. The reduction of 0.0415 eV has a 95% formula-clusterbootstrap interval of 0.0255 to 0.0578 eV. Five formula-group splits,including the primary split, give mean errors of 0.6315 and 0.6581eV, respectively. These results support periodic density overlap descriptorsfor modeling computed band gaps when valence charge densities areavailable.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":"22 21","pages":"11410–11418"},"PeriodicalIF":5.8,"publicationDate":"2026-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149251103","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":"r2SCAN-Based Composite Methods for Solid-State Calculations","authors":"Mattia Raimondo,&nbsp;Lorenzo Donà,&nbsp;Bartolomeo Civalleri","doi":"10.1021/acs.jctc.6c00852","DOIUrl":"10.1021/acs.jctc.6c00852","url":null,"abstract":"<p>In this work, wehave extended the existing “sol-3c”composite methods to meta-GGA XC functionals using the r²SCAN one. The corresponding r²SCAN-based composite methodsin their pure and hybrid HF/DFT variants have been developed in combinationwith double-ζ (sol-def2-mSVP) and triple-ζ (pob-TZVP-rev2)quality basis sets specifically adapted to solids. The r²SCAN “sol-3c/pob-3c” methods are validated and testedon benchmark sets containing both intermolecular adducts (S66 ×7), molecular crystals (X23, DMC8) and simple inorganic solids (SS20).They are then applied to selected applications of relevance for solid-statecalculations such as the study of the energetic and structural propertiesof layered materials, the mechanical, electronic and vibrational propertiesof α-quartz, relative stability of all-silica zeolite polymorphs,α-quartz vs stishovite phase transition and the adsorption ofsmall molecules on the surface of inorganic solids and in the poreof metal–organic frameworks. The new methods show broad applicabilitywith accurate results and an adjustable and affordable computationalcost.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":"22 21","pages":"11251–11265"},"PeriodicalIF":5.8,"publicationDate":"2026-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148858533","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":"DeepHartree: A Poisson-Coupled Neural Field for One-Shot Density Functional Theory","authors":"Jiankun Wu,&nbsp;Jinming Fan,&nbsp;Chao Qian,&nbsp;Shaodong Zhou","doi":"10.1021/acs.jctc.6c01494","DOIUrl":"10.1021/acs.jctc.6c01494","url":null,"abstract":"<p>Linear-combination-of-atomic-orbital (LCAO) density functionaltheory (DFT) incurs steep costs when it constructs Coulomb terms anditerates the self-consistent field (SCF) equations. Matrix-learningapproaches can bypass parts of this workflow, but their outputs inheritthe dimensions and conventions of a fixed orbital basis. We introduceDeepHartree, a Poisson-coupled neural field that connects continuousreal-space prediction to LCAO electronic structure. An <em>E</em>(3)-equivariant network predicts the Hartree potential, and the Poissonequation converts this potential into electron density. Atom-centeredGaussian fields resolve the near-nuclear region, while a molecule-levelcorrection enforces the electron count. Analytical Gaussian integralsand numerical quadrature connect the predicted fields to the Kohn–Shammatrix. One diagonalization provides estimates of frontier orbitalenergies and the occupied subspace, and yields the density matrixused to evaluate energy components and initialize SCF. Mean molecularweighted normalized mean absolute errors (wNMAEs) are 0.110% for QM9and 0.199% for VQM24 on PySCF level-0 grids, increasing to 1.366%and 2.218%, respectively, on level-3 grids. Models trained on def2-SVPdensities transfer without retraining to def2-TZVP and def2-QZVP,retaining total-energy accuracy and occupied-subspace agreement. TheHybrid method combines finite differences and first-order automaticdifferentiation with analytical Gaussian integrals and multilevelquadrature to accelerate field-to-matrix conversion. It achieves near-lineardensity-matrix construction in both alkanes and PEG chains. For the1004-atom PEG chain, this construction is 8.75-fold faster than fullAD and 7.34-fold faster than converged GPU4PySCF SCF. Without fine-tuning,the QM9 model attains a total-energy MAE of 9.437 meV atom^–1 on 1000 larger OE62 molecules. Across the three basis sets, DeepHartreeinitialization reduces mean QM9 SCF iteration counts by 15.3–15.6%.For a 620-atom organic cage, it reduces total SCF wall-clock timeby 12.9%, including initialization. These results establish continuouselectrostatic fields as an accurate and scalable interface betweenmachine learning and LCAO DFT.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":"22 21","pages":"11393–11409"},"PeriodicalIF":5.8,"publicationDate":"2026-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149229970","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":"P3MaZe: a Mass-Zero Constrained-Dynamics Formulation of Particle–Mesh Electrostatics","authors":"Federica Troni,&nbsp;Violette Gontran,&nbsp;Davide Grassano,&nbsp;Sara Bonella","doi":"10.1021/acs.jctc.6c01315","DOIUrl":"10.1021/acs.jctc.6c01315","url":null,"abstract":"<p>We introduce P3MaZe, a real-space particle–meshelectrostaticmethod that combines the standard short-range/long-range decompositionof Particle–Particle Particle–Mesh (P3M) electrostaticswith the Mass-Zero constrained dynamics (MaZe) framework. In thisformulation, the smooth long-range electrostatic potential is representedon a mesh as a zero-inertia auxiliary field, while the discretizedPoisson equation is enforced as a holonomic constraint during moleculardynamics. By retaining the standard P3M decomposition, P3MaZe preservesthe systematic accuracy controls associated with the real-space cutoff,the Ewald splitting, the mesh spacing, and the charge assignment procedure,while replacing the conventional multigrid Poisson solver by a constrainedcorrection problem. The method is validated for molten NaCl and simplepoint-charge flexible water (SPC/Fw). Values of the potential, structural,translational, collective, and rotational dynamical observables arein quantitative agreement with those obtained with established electrostaticmethods, including real-space P3M, and Ewald summation. We furthershow that, for the linear, time-independent Poisson operator consideredhere, the constrained correction problem is algebraically equivalentto a direct Poisson solve initialized by a chronological extrapolationof the mesh potential, a connection not previously established forMaZe-based approaches. This result identifies the precise mechanismthrough which acceleration of the iterative long-range solve is achieved,and we exploit it systematically through higher-order predictors tofurther improve convergence, while retaining the expected linear scalingwith system size. Because the underlying extrapolation is independentof the constrained-dynamics formalism, it can, of course, also beused directly to accelerate conventional iterative real-space Poissonsolvers.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":"22 21","pages":"11442–11455"},"PeriodicalIF":5.8,"publicationDate":"2026-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149248650","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":"FIREX: Fast-Ion Replica EXchange Enables Rapid Sampling of Mg2+ Pairing with Nucleic Acids","authors":"Laetitia Kantin,&nbsp;Julie Puyo-Fourtine,&nbsp;Jérôme Hénin,&nbsp;Elise Duboué Dijon","doi":"10.1021/acs.jctc.6c01331","DOIUrl":"10.1021/acs.jctc.6c01331","url":null,"abstract":"<p>We introduce Fast-Ion Replica EXchange (FIREX), a HamiltonianReplicaExchange-based sampling strategy specifically designed to enhancethe sampling of metastable divalent cation binding with phosphategroups, e.g., in nucleic acids. We parametrize the scheme and demonstrateits convergence and efficiency on a model DNA dodecamer with Mg²⁺ ions, using two force fields with very different ion-pairingproperties to test its robustness. The highest replica is driven bya “fast-ion” Hamiltonian combining a reduced ion chargewith a tuned pair-specific Lennard-Jones interaction between the ionand phosphate oxygen atoms, yielding rapid exchange as well as moderatebinding affinity, amenable to efficient sampling. By enabling therobust quantification of divalent ion interactions in simulations,FIREX could be instrumental in refining force fields to better captureion–molecule interactions. The provided guidelines for thedesign of fast-ion models suitable for FIREX are general and couldbe extended to other force fields or types of biomolecular moieties,such as carboxylated protein side chains, phosphorylated proteins,or other interaction sites on nucleic acids. We further demonstratehow FIREX can be efficiently combined with the REST2 method to achievesimultaneous sampling of ion binding and biomolecular conformationswith the Amber-OL15-ECC force field, while convergence remains challengingwith the original Amber-OL15. This shows that different HREX schemes,including REST2, can be merged into a combined, highly cost-effectiveHREX scheme and is a step toward robust, computationally efficientsampling of charged biomolecules in interaction with divalent ions.</p>","PeriodicalId":45,"journal":{"name":"Journal of Chemical Theory and Computation","volume":"22 21","pages":"11647–11661"},"PeriodicalIF":5.8,"publicationDate":"2026-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149251036","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}