Journal of Computational Chemistry最新文献

The evaluation of oxidation and reduction potentials is a pivotal task in various chemical fields. However, their accurate prediction by theoretical computations, which is a complementary task and sometimes the only alternative to experimental measurement, may be often resource-intensive and time-consuming. This paper addresses this challenge through the application of machine learning techniques, with a particular focus on graph-based methods (such as graph edit distances, graph kernels, and graph neural networks) that are reviewed to enlighten their deep links with theoretical chemistry. To this aim, we establish the ORedOx159 database, a comprehensive, homogeneous (with reference values stemming from density functional theory calculations), and reliable resource containing 318 one-electron reduction and oxidation reactions and featuring 159 large organic compounds. Subsequently, we provide an instructive overview of the good practice in machine learning and of commonly utilized machine learning models. We then assess their predictive performances on the ORedOx159 dataset through extensive analyses. Our simulations using descriptors that are computed in an almost instantaneous way result in a notable improvement in prediction accuracy, with mean absolute error (MAE) values equal to 5.6 kcal mol$��{}^{�-�1�}�$ for reduction and 7.2 kcal mol$��{}^{�-�1�}�$ for oxidation potentials, which paves a way toward efficient in silico design of new electrochemical systems.

This study delves into the nature of individual hydrogen bonds and the relationship between metal cations and hydrogen bonding in the Watson–Crick guanine–cytosine (GC) base pair and its alkali and alkaline earth cation-containing complexes (Mⁿ⁺–GC). The findings reveal how metal cations affect the nature and strength of individual hydrogen bonds. The study employs interacting quantum atoms (IQA) analysis to comprehensively understand three individual hydrogen bonds within the GC base pair and its cationic derivatives. These analyses unveil the nature and strength of hydrogen bonds and serve as a valuable reference for exploring the impact of cations (and other factors) on each hydrogen bond. All the H$��\cdots �$D interactions (H is hydrogen and D is oxygen or nitrogen) in the GC base pair are primarily electrostatic in nature, with the charge transfer component playing a substantial role. Introducing a metal cation perturbs all H$��\cdots �$D interatomic interactions in the system, weakening the nearest hydrogen bond to the cation (indicated by a) and reinforcing the other (b and c) interactions. Notably, the interaction a, the strongest H$��\cdots �$D interaction in the GC base pair, becomes the weakest in the Mⁿ⁺–GC complexes. A broader perspective on the stability of GC and Mⁿ⁺–GC complexes is provided through interacting quantum fragments (IQF) analysis. This approach considers all pairwise interactions between fragments and intra-fragment components, offering a complete view of the factors that stabilize and destabilize GC and Mⁿ⁺–GC complexes. The IQF analysis underscores the importance of electron sharing, with the dominant contribution arising from the inter-fragment exchange-correlation term, in shaping and sustaining GC and Mⁿ⁺–GC complexes. From this point of view, alkaline and alkaline earth cations have distinct effects, with alkaline cations generally weakening inter-fragment interactions and alkaline earth cations strengthening them. In addition, IQA and IQF calculations demonstrate that the hydration of cations led to small changes in the hydrogen bonding network. Finally, the IQA interatomic energies associated with the hydrogen bonds and also inter-fragment interaction energies provide robust indicators for characterizing hydrogen bonds and complex stability, showing a strong correlation with total interaction energies.

The development of infrared difference spectroscopy provides unprecedented insights on structures of complex molecules like metalloproteins. However, the relevant information can be hard to find among the many bands of the vibrational spectra. The ab initio modeling is very helpful to assign the frequencies to vibrational modes but it is a challenge to process the huge quantity of data into descriptors useful for experimentalists. To this end, we developed a new tool called VIBMOL allowing to analyze vibrational modes of molecules from hessian matrices calculated with common quantum chemistry codes. VIBMOL program runs on Unix machines. Through a new graphical interface, the users can calculate the normal modes of molecules, visualize them, simulate infrared spectra, and explore the Potential Energy Distribution of normal modes among any set of vibration coordinates. It is combined with an interface program (gosdmu) formatting relevant data from the GAUSSIAN program. VIBMOL code is available upon request to the authors. A discussion is provided to help the readers to choose between a large choice of different software and it shows how VIBMOL can make the IR assignment easier in the context of collaborations with experimentalists.

In this work, we discuss the use of a recently introduced machine learning (ML) technique known as Fourier neural operators (FNO) as an efficient alternative to the traditional solution of the time-dependent Schrödinger equation (TDSE). FNOs are ML models which are employed in the approximated solution of partial differential equations. For a wavepacket propagating in an anharmonic potential and for a tunneling system, we show that the FNO approach can accurately and faithfully model wavepacket propagation via the density. Additionally, we demonstrate that FNOs can be a suitable replacement for traditional TDSE solvers in cases where the results of the quantum dynamical simulation are required repeatedly such as in the case of parameter optimization problems (e.g., control). The speed-up from the FNO method allows for its combination with the Markov-chain Monte Carlo approach in applications that involve solving inverse problems such as optimal and coherent laser control of the outcome of dynamical processes.

Classical scoring functions may exhibit low accuracy in determining ligand binding affinity for proteins. The availability of both protein–ligand structures and affinity data make it possible to develop machine-learning models focused on specific protein systems with superior predictive performance. Here, we report a new methodology named SAnDReS that combines AutoDock Vina 1.2 with 54 regression methods available in Scikit-Learn to calculate binding affinity based on protein–ligand structures. This approach allows exploration of the scoring function space. SAnDReS generates machine-learning models based on crystal, docked, and AlphaFold-generated structures. As a proof of concept, we examine the performance of SAnDReS-generated models in three case studies. For all three cases, our models outperformed classical scoring functions. Also, SAnDReS-generated models showed predictive performance close to or better than other machine-learning models such as K_DEEP, CSM-lig, and Δ_VinaRF₂₀. SAnDReS 2.0 is available to download at https://github.com/azevedolab/sandres.

Acetonitrile, a polar molecule that cannot form hydrogen bonds on its own, interacts with solvent molecules mainly through the lone pair of its nitrogen atom and the π electrons of its CN triple bond [Correction added on 17 July 2024, after first online publication: Acetole has been changed to Acetonitrile in the preceeding sentence.]. Interestingly, acetonitrile exhibits an unexpected strengthening of the triple bond's force constant in an aqueous environment, leading to an upshift (blueshift) in the corresponding stretching vibration: this effect contrasts with the usual consequence of hydrogen bonding on the vibrational frequencies of the acceptor groups, that is, frequency redshift. This investigation elucidates this phenomenon using Raman spectroscopy to examine the behavior of acetonitrile in organic solvent, water, and silver ion aqueous solutions, where an even more pronounced upshift is observed. Raman spectroscopy is particularly well suited for analyzing aqueous solutions due to the minimal scattering effect of water molecules across most of the vibrational spectrum. Computational approaches, both static and dynamical, based on Density Functional Theory and hybrid functionals, are employed here to interpret these findings, and accurately reproduce the vibrational frequencies of acetonitrile in different environments. Our calculations also allow an explanation for this unique behavior in terms of electric charge displacements. On the other hand, the study of the interaction of acetonitrile with water molecules and metal ions is relevant for the use of this molecule as a solvent in both chemical and pharmaceutical applications.

Observational data show complex organic molecules in the interstellar medium (ISM). Hydrogenation of small unsaturated carbon double bond could be one way for molecular complexification. It is important to understand how such reactivity occurs in the very cold and low-pressure ISM. Yet, there is water ice in the ISM, either as grain or as mantle around grains. Therefore, the addition of atomic hydrogen on double-bonded carbon in a series of seven molecules have been studied and it was found that water catalyzes this reaction. The origin of the catalysis is a weak charge transfer between the π MO of the unsaturated molecule and H atom, allowing a stabilizing interaction with H₂O. This mechanism is rationalized using the non-covalent interaction and the quantum theory of atoms in molecules approaches.

Due to the potential applications in next-generation micro/nano electronic devices and functional materials, magnetic germanium (Ge)-based clusters are receiving increasing attention. In this work, we reported the structures, electronic and magnetic properties of CrMnGe_n with sizes n = 3–20. Transition metals (TMs) of Cr and Mn tend to stay together and be surrounded by Ge atoms. Small sized clusters with n ≤ 8 prefer to adopt bipyramid-based structures as the motifs with the excess Ge atoms absorbed on the surface. Starting from n = 9, the structure with one TM atom interior appears and persists until n = 16, and for larger sizes n = 17–20, the two TM atoms are full-encapsulated by Ge atoms to form endohedral structures. The Hirshfeld population analyses show that Cr atom always acts as the electron donor, while Mn atom is always the acceptor except for sizes 3 and 6. The average binding energies of these clusters increase with cluster size n, sharing a very similar trend as that of CrMnSi_n (n = 4–20) clusters, which indicates that it is favorable to form large-sized clusters. CrMnGe_n (n = 6, 13, 16, 19, and 20) clusters prefer to exhibit ferromagnetic Cr–Mn coupling, while the remaining clusters are ferrimagnetic.

Here, TS-tools is presented, a Python package facilitating the automated localization of transition states (TS) based on a textual reaction SMILES input. TS searches can either be performed at xTB or DFT level of theory, with the former yielding guesses at marginal computational cost, and the latter directly yielding accurate structures at greater expense. On a benchmarking dataset of mono- and bimolecular reactions, TS-tools reaches an excellent success rate of 95% already at xTB level of theory. For tri- and multimolecular reaction pathways - which are typically not benchmarked when developing new automated TS search approaches, yet are relevant for various types of reactivity, cf. solvent- and autocatalysis and enzymatic reactivity - TS-tools retains its ability to identify TS geometries, though a DFT treatment becomes essential in many cases. Throughout the presented applications, a particular emphasis is placed on solvation-induced mechanistic changes, another issue that received limited attention in the automated TS search literature so far.