This work employs a combined fast marching tree (FMT)-string algorithm to determine the minimum energy path (MEP). The idea behind this method is that using FMT can give the MEP good initial approximations, so only a few iterations are required for refinement. This saves time and the number of iterations required compared to using a linear-interpolated (LI) initial guess. After that, we show that this method works in a number of model free energy surfaces (FES), including the well-known Muller-Brown potential, where FMT-string is better than the LI counterpart at getting a converged MEP. We also explore the potential applications of this method in molecular systems.

Excited state calculations were performed on fourth-position indole derivatives to evaluate substituent effects on indole's absorption. Results show correlation of excitation energies to the electrophilicity parameter, and indicate that theory can predict excited state properties of indole derivatives with electron-withdrawing substituents.

Molecular and ionic isomers of complexes in Cl₂-LB systems (LB = nitrogen-containing Lewis base) were explored at M06-2X/def2-TZVPD level of theory, verified against benchmark computations at CCSD(T) ab initio level for selected compounds. Formation of contact ion pairs [Cl∙(LB)₂]⁺…C${\mathrm{l}}_3^{-}$ makes the heterogeneous halogen-halogen bond splitting both energetically and kinetically favorable.

Target-specific scoring function for human dihydroorotate dehydrogenase (TSSF-hDHODH) is a machine learning-based target-specific scoring function based on molecular docking structures for predicting the binding affinities of inhibitors to hDHODH. This high-performing scoring function is developed by optimally combining interaction and ligand features with five machine learning algorithms and can be applied to virtual screening of potential inhibitors from compound libraries. This study can aid in the development of drugs for hDHODH and scoring functions for other targets.

An theoretical study on the luminescent and reaction mechanism of dansyl-based fluorescence probe for detecting hydrogen sulfide is conducted and the results can explain the “turn-on” effect of fluorescent probe in the presence of H₂S.

In the protonation of MX₂ and M₂X₂ (M = Be, Mg; X = H, H) complexes with nitrogen bases, the hydrogen or fluoride abstraction process competes with the protonation at the base. The abstraction process is consistently favored for hydrides.

Detailed noncovalent bonding analyses unveil the existence of a fluxional halogen bonds (FXBs) in a series of linear (IC₆F₄I)_m(OONC₆H₄NOO)_n (m + n = 5) complexes which fluctuates reversibly in a linear $\mathrm{C}<span\; class="icomoon"></span>\mathrm{I}\cdots \mathrm{O}\left(\mathrm{GS}\right)\rightleftharpoons \text{bifurcated}\mathrm{C}<span\; class="icomoon"></span>\mathrm{I}<{\mathrm{O}}_2\mathrm{N}\left(\mathrm{TS}\right)\rightleftharpoons \text{linear}\mathrm{C}<span\; class="icomoon"></span>\mathrm{I}\cdots \mathrm{O}\left({\mathrm{GS}}^{\text{'}}\right)$ mechanism, facilitating the fluxionality of the systems.

A new biaryl monophosphine ligand designed by theoretical calculations with negative electrostatic potential close to the palladium-bonded fluoride ion is predicted to increase the reaction rate and regioselectivity in palladium-catalyzed aromatic fluorination.

Open-source protocol to estimate ligand strain after docking: two quantum mechanics (QM) single-point energy calculations with a Monte Carlo (MC) based ligand minimization in-between. The MC simulation uses the ANI-2x (QM approximating) force field and is performed in the dihedral space.

Due to their potential for eco-friendly energetic applications, [FeFe] hydrogenase models attract interest of both experimental and computational chemists. Here, a series of DFT methods was tested to establish their performance in describing the molecular and crystal structures of recently synthesized [FeFe] hydrogenase models. r²SCAN is recommended for predicting the structure of such models with a high degree of accuracy.

Fragment-based de novo design can construct new molecules from scratch from an input library of allowable fragments and torsion types. In this work, a new series of biasing methods have been developed for the program DOCK6, which facilitate the construction of molecular ensembles containing fragment and torsion populations in proportion to those seen in drug-like libraries. Changes to clustering and pruning have also been made to increase the overall efficiency of the methods, producing many more unique molecules in a similar amount of time.

Self-supervised pre-training pipeline integrated with high-throughput virtual screening for non-nucleoside reverse transcriptase inhibitors (NNRTIs).

Small molecule conformational sampling is crucial for exploring the vast conformational space within specific chemical environments. Recent advancements have led to the emergence of various conformational sampling methods. This study evaluates the impact of different sampling methods on molecular docking. Among the methods, the CCDC Conformer Generator demonstrates superior bioactive pose reproducibility and comparable screening. Additionally, each sampling method excels in specific cases, and combining various methods can enhance the overall performance.

The cyclic nitrenium ions with seven membered rings are aromatic but not very stable. Their electronic, energetic properties can be fine-tuned and the species carrying ΔE_S-T ≈ 50 kcal/mol may prove to be stable.

Acetanilide and paracetamol photophysics after the photoexcitation to the second singlet excited state, which can lead to the formation of the photo-fries photoproduct.

Multiconfiguration density-coherence functional theory is a low-cost and high-accuracy electronic structure method that recovers dynamic correlation in inherently multiconfigurational molecules and transition states by expressing the nonclassical energy as a functional of the unpaired density computed from the one-particle reduced density matrix in the coordinate representation. Here we propose a new, more physical expression for the unpaired density and use it to optimize a density coherence functional with improved physical interpretation and upgraded accuracy.

In this work, we investigated the structures and electronic and magnetic properties of Cr₂Pb_n (n = 3–20) clusters based on a self-developed genetic algorithm coupled with density functional theory calculations. It found that all Cr₂Pb_n clusters are antiferromagnets, except for the ferrimagnetic Cr₂Pb₁₁ with a net magnetic moment of 2 μ_B. The discovered stable Cr₂Pb₁₇ cluster can assemble into dimers and trimers while maintaining its geometric structure and antiferromagnetic properties, indicating the potential of becoming a structural unit for cluster-assembled antiferromagnetic materials.

MolAR is the first memory-safe library for analysis of MD simulations written in Rust. MolAR is intended to explore the advantages and challenges of implementing molecular analysis software in the memory-safe natively compiled language and to develop specific memory-safe abstractions for this kind of software. MolAR outperforms popular molecular analysis libraries and tools, which makes it attractive for implementing computationally intensive analysis tasks.

The magnetic exchange coupling constant calculation for azulene-bridged diradicals is highly sensitive to the amount of Hartree–Fock (HF) exchange in the density functionals. The hybrid functionals accurately determine the sign of ferromagnetic coupling. In a particular value of HF exchange in the hybrid functionals, they predict the ferro- and antiferromagnetic coupling correctly.

Global optimization was performed on the BXY₃ (X = B, Al, Ga; Y = C, Si, Ge) clusters and neutral planar tetracoordinated boron containing global minimum structures with 18 valence electrons, were obtained for most of the systems. The ptB global minimum structures are stable at 500 K temperature and the aromaticity analysis revealed both σ- and π-electron delocalization.

The local mode decomposition revealed that the CH₂ rocking mode, associated with the CSH bending, appears only for the gauche isomer in accordance with the experimental assignments. The most stable dimer is the trans(right)–gauche(left) structure. The HOMO orbital is located on the donor monomer, and the LUMO orbital is located on the acceptor monomer, supporting the largest redshift of 33 cm⁻¹ obtained for the gauche isomer. The configuration shows the S–H donor directed to the S acceptor with the NCI isosurface. The LUMO orbital also shows the S⋅⋅⋅S interatomic distance. Dispersion is the dominant term in this dimer, while electrostatic interactions comprise 83% of dispersion. ELF values at the critical point of S–H⋅⋅⋅S interaction describe an almost linear form of the interaction energy.

QM/MM MP2 calculations bring new insights into the catalytic mechanism of ampicillin hydrolysis by New Delhi Metallo-β-lactamase 1 (NDM-1). It is found that the rate-determining step is the dissociation of the hydrolyzed ampicillin via a solvent proton transfer to the newly form carboxylate group.

Nanoparticles were developed in vitro and in silico to overcome limitations of cisplatin (CDDP), Platinum (Pt) based anticancer drug. Citrate is commonly biocompatible in nano-drug synthesis. In this study, density functional theory (DFT) has been investigated to study the physicochemical and electronic properties of six CDDP-citrate complexes in water and DMSO. The frontier orbitals, electron densities mapping, and electrostatic potential analysis sheds light on the effective complex for nanomedicine applications.

CGPDTA uses transfer learning to predict drug-target binding affinity by extracting effective drug-protein interaction information through the lens of a drug substructure graph.

This study presents a novel computational protocol for folding guanine quadruplexes (GQs) from single-strand conformation to their native state. Utilizing two enhanced sampling techniques, the protocol models multichannel folding pathways, elucidating diverse intermediates and mechanisms. Despite its focus on the all-anti parallel GQ topology, this approach significantly enhances our understanding of GQ stability and dynamics and can be adapted for more complex folding systems.

This paper reports the improvement in the efficiency of Embedded Cluster Model (ECM) calculations in ORCA thanks to the implementation of the Fast Multipole Method. The outcome of the work is a one order of magnitude minimum speedup for the evaluation of the electrostatic potentials.

The acid dissociation constant (pK_a) plays a crucial role in applications ranging from pharmaceutical synthesis to environmental science. This work demonstrates that popular solvation model COSMO-RS can be used in conjunction with aqueous pK_a data to predict the pK_a in many non-aqueous solvents, which can then be leveraged to approximate the transfer free energy of the proton between solvents. This work also explores inconsistencies in experimental data that sometimes lead to worse model agreement.

Increase of the solvent polarity leads to changes in the chemical bonding from molecular complexes to 3-center 4-electron bonds and to the contact ion pairs, facilitating heterolytic halogen-halogen bond splitting. Structural changes upon solvation in polar solvents are significant and single point solvation energy computations on gas phase optimized geometries are not reliable for a such systems.

Using nonadiabatic dynamics, we uncover an ultrafast hydrogen migration-driven internal conversion pathway in diketopyrrolopyrrole dimers with a sub-picosecond excited-state lifetime. This work highlights the critical impact of state intersections on photophysical performance, offering insights for functionalizing DPP systems to suppress nonradiative decay.

We investigate, using DFT, cobalt-coordinated tetracyanoquinodimethane (R-CoTCNQ) as a metal–organic framework catalysis for the CO₂ electroreduction reaction. We find that reduced (charged) R-CoTCNQ may produce formic acid and/or formaldehyde reaction products as both a supported and unsupported catalyst.

Groupy is a package targeted for calculating different molecular properties such as density, normal boiling point, normal melting point, critical temperature, flash point, critical pressure, critical volume, Gibbs free energy, enthalpy of formation, enthalpy of fusion, and enthalpy of vaporization. Groupy also can serve as an intermediary platform that can create input files for Gaussian and visualize 3D structure of a molecule based on SMILES.

All-benzene catenanes experience faster relaxations than individual units. They present overlapping energy manifolds that include electronic excited states spatially localized on the different moieties, increasing the density of states. This result suggests a viable strategy for tuning the internal conversion rates in a quest for their utilization for new applications.

12348 measured transitions in the wavenumber range 578–9318 cm⁻¹ for carbon dioxide (isotopologue 638) were used° to calculating 3975 empirical rovibrational energy levels with J° up to 82.

Folates, essential cofactors in one-carbon metabolism, play a pivotal role in critical physiological processes, including biosynthesis, epigenetic regulation, and redox defense. This study presents a molecular mechanics model for folates, parameterized for compatibility with the CHARMM36 all-atom force field. The model includes 18 folate derivatives across multiple redox and protonation states, employing quantum mechanical reference data to derive partial charges and optimize bonded terms. Simulations of protein complexes validated the improved accuracy of the new parameters in capturing folate–protein interactions. This model allows studying folate dynamics and serves as a foundation for further parameterization of folate derivatives.

The F12-TcCR+TZ QFF alleviates many of the issues present in other QFF + VPT2 anharmonic frequency computations for aluminum-bearing molecules. This allows for more accurate computational spectroscopic data for atmospheric and interstellar detection of these aluminum-bearing molecules.

Including mutual polarization effects in quantum mechanics/molecular mechanics (QM/MM) approaches is important especially when excited states are involved. Here we present a polarizable embedding QM/MM model that couples Complete Active Space Self-Consistent Field (CASSCF) descriptions and the AMOEBA polarizable force field. The method is tested on two photoreceptor proteins for which we simulate the excitation process through geometry optimization of the ground state and calculation of the vertical excitation energy

Numerical schemes are developed to compute the five solvation free energy components in terms of one electrostatic potential, $$$ u, $$$ and $$$ n $$$ ionic concentration functions $$$ , \kern0.5em {\left\{{c}_i\right\}}_{i=1}^n $$$, calculated by our nuSMPBIC, SMPBIC, and PBIC finite element packages. The schemes are then integrated with our improved mesh generation package, the PDB2PQR package for generating PQR files, and the OPM database for downloading PDB files, resulting in the voltage-dependent anion channel solvation free energy calculation package.

We investigate the structural, electronic, optical, and elastic properties of Cs₂NaScX₆ (X = Cl, Br) double perovskites using DFT in WIEN2k. The compounds exhibit cubic symmetry, mechanical stability, ductile behavior, and pronounced visible-light absorption, with band gaps of 3.138 eV and 3.977 eV, respectively. These findings highlight their potential for advanced photovoltaic applications.

Deuterium isotope effects on interaction energies and geometrical parameters in ionic H₃O⁺…π interactions were comprehensively analyzed using multi-component quantum mechanics. Replacing H₃O⁺ with D₃O⁺ reduced interaction energies and induced changes in hydrogen-bonded distances. Natural energy decomposition analysis revealed strong correlations between H/D isotope effects on these distances and energy components.

As can be seen in the above box plot, the counterpoise correction on the post-CCSD(T) corrections is over an order of magnitude smaller than that on the connected triples, and two orders smaller than that on the CCSD correlation energy.

Calculated resonance Raman spectra (647 nm excitation) for a Type 1 (T1) copper site, comparing simplest and most complex models. The simplest model exhibits one strong band, while the complex model shows band splitting, demonstrating how increased model sophistication affects predicted vibrational spectra of copper proteins.

Advanced machine learning methods could be employed to model key properties, including electronic energy and vibration frequencies, of the Lithium-ion battery electrolytes.

We present a delta machine learning (Δ-ML) tool that corrects simplified Tamm–Dancoff Approximation excited state energies to produce second order algebraic diagrammatic construction (ADC(2))-like results. We use Δ-ML to screen aza-triangulenes to find those with inverted singlet-to-triplet-gaps (INVEST) and have excited states suitable for photocatalytic water splitting (PWS). Our Δ-ML model has an 85% recall for systems identified by ADC(2) as having both the INVEST and PWS properties, producing MADs within 28–50 meV of ADC(2).

Using the QM/MM method, we analyzed the strength of FeS bonds in heme groups of oxidized and reduced forms of bacterioferritin. The strength of FeS bonds was correlated with bond length, energy density at bond critical point, and charge difference between F and S atoms. The change of oxidation state from ferrous to ferric, makes FeS bonds weaker, longer, more covalent, and more polar. We have also found that stronger FeS bonds correlate with stronger SFeS bond bending.

Local constraint violations by the BP86 functional, quantified by our metric, trending with total energy errors: molecules with more violations also have higher errors in DFT total energies.

Thermodynamic stability in transition metal-containing diatomic dication—VH²⁺.

The individual hydrogen bond energies in molecular crystal of nitromalonamide and salicylic acid are calculated by the MTA-based and the Frags-in-Frags methods. The direct application of MTA-based method is very expensive and may not be possible with off-the-shelf-hardware. The Frags-in-Frags method provides accurate results and is inexpensive.

A neural net has been trained to fit conformer energy differences obtained from large basis set density functional calculations. Applied as a correction to the MMFF molecular model, this approach leads to dramatic improvement with a small (insignificant) increase in calculation cost.