Browse
Submit Data
Databases
API
Help

Dataset Information

10 Views

0 Connections

0 Citations

0 Reanalyses

0 Downloads

Omics score: 0

Regio-selectivity prediction with a machine-learned reaction representation and on-the-fly quantum mechanical descriptors.

ABSTRACT: Accurate and rapid evaluation of whether substrates can undergo the desired the transformation is crucial and challenging for both human knowledge and computer predictions. Despite the potential of machine learning in predicting chemical reactivity such as selectivity, popular feature engineering and learning methods are either time-consuming or data-hungry. We introduce a new method that combines machine-learned reaction representation with selected quantum mechanical descriptors to predict regio-selectivity in general substitution reactions. We construct a reactivity descriptor database based on ab initio calculations of 130k organic molecules, and train a multi-task constrained model to calculate demanded descriptors on-the-fly. The proposed platform enhances the inter/extra-polated performance for regio-selectivity predictions and enables learning from small datasets with just hundreds of examples. Furthermore, the proposed protocol is demonstrated to be generally applicable to a diverse range of chemical spaces. For three general types of substitution reactions (aromatic C-H functionalization, aromatic C-X substitution, and other substitution reactions) curated from a commercial database, the fusion model achieves 89.7%, 96.7%, and 97.2% top-1 accuracy in predicting the major outcome, respectively, each using 5000 training reactions. Using predicted descriptors, the fusion model is end-to-end, and requires approximately only 70 ms per reaction to predict the selectivity from reaction SMILES strings.

SUBMITTER: Guan Y

PROVIDER: S-EPMC8179287 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Json Xml

Similar Datasets

Mechanistic Insights into Enzyme Catalysis from Explaining Machine-Learned Quantum Mechanical and Molecular Mechanical Minimum Energy Pathways.

Project description:With the increasing popularity of machine learning (ML) applications, the demand for explainable artificial intelligence techniques to explain ML models developed for computational chemistry has also emerged. In this study, we present the development of the Boltzmann-weighted cumulative integrated gradients (BCIG) approach for effective explanation of mechanistic insights into ML models trained on high-level quantum mechanical and molecular mechanical (QM/MM) minimum energy pathways. Using the acylation reactions of the Toho-1 β-lactamase and two antibiotics (ampicillin and cefalexin) as the model systems, we show that the BCIG approach could quantitatively attribute the energetic contribution in one system and the relative reactivity of individual steps across different systems to specific chemical processes such as the bond making/breaking and proton transfers. The proposed BCIG contribution attribution method quantifies chemistry-interpretable insights in terms of contributions from each elementary chemical process, which is in agreement with the validating QM/MM calculations and our intuitive mechanistic understandings of the model reactions.

| S-EPMC9344433 | biostudies-literature

Machine Learning of Reaction Properties via Learned Representations of the Condensed Graph of Reaction.

Project description:The estimation of chemical reaction properties such as activation energies, rates, or yields is a central topic of computational chemistry. In contrast to molecular properties, where machine learning approaches such as graph convolutional neural networks (GCNNs) have excelled for a wide variety of tasks, no general and transferable adaptations of GCNNs for reactions have been developed yet. We therefore combined a popular cheminformatics reaction representation, the so-called condensed graph of reaction (CGR), with a recent GCNN architecture to arrive at a versatile, robust, and compact deep learning model. The CGR is a superposition of the reactant and product graphs of a chemical reaction and thus an ideal input for graph-based machine learning approaches. The model learns to create a data-driven, task-dependent reaction embedding that does not rely on expert knowledge, similar to current molecular GCNNs. Our approach outperforms current state-of-the-art models in accuracy, is applicable even to imbalanced reactions, and possesses excellent predictive capabilities for diverse target properties, such as activation energies, reaction enthalpies, rate constants, yields, or reaction classes. We furthermore curated a large set of atom-mapped reactions along with their target properties, which can serve as benchmark data sets for future work. All data sets and the developed reaction GCNN model are available online, free of charge, and open source.

| S-EPMC9092344 | biostudies-literature

The quantum-mechanical Coulomb propagator in an L<sup>2</sup> function representation.

Project description:The quantum-mechanical Coulomb propagator is represented in a square-integrable basis of Sturmian functions. Herein, the Stieltjes integral containing the Coulomb spectral function as a weight is evaluated. The Coulomb propagator generally consists of two parts. The sum of the discrete part of the spectrum is extrapolated numerically, while three integration procedures are applied to the continuum part of the oscillating integral: the Gauss-Pollaczek quadrature, the Gauss-Legendre quadrature along the real axis, and a transformation into a contour integral in the complex plane with the subsequent Gauss-Legendre quadrature. Using the contour integral, the Coulomb propagator can be calculated very accurately from an L[Formula: see text] basis. Using the three-term recursion relation of the Pollaczek polynomials, an effective algorithm is herein presented to reduce the number of integrations. Numerical results are presented and discussed for all procedures.

| S-EPMC8460653 | biostudies-literature

Combining machine learning and quantum mechanics yields more chemically aware molecular descriptors for medicinal chemistry applications.

Project description:Molecular interaction fields (MIFs), describing molecules in terms of their ability to interact with any chemical entity, are one of the most established and versatile concepts in drug discovery. Improvement of this molecular description is highly desirable for in silico drug discovery and medicinal chemistry applications. In this work, we revised a well-established molecular mechanics' force field and applied a hybrid quantum mechanics and machine learning approach to parametrize the hydrogen-bonding (HB) potentials of small molecules, improving this aspect of the molecular description. Approximately 66,000 molecules were chosen from available drug databases and subjected to density functional theory calculations (DFT). For each atom, the molecular electrostatic potential (EP) was extracted and used to derive new HB energy contributions; this was subsequently combined with a fingerprint-based description of the structural environment via partial least squares modeling, enabling the new potentials to be used for molecules outside of the training set. We demonstrate that parameter prediction for molecules outside of the training set correlates with their DFT-derived EP, and that there is correlation of the new potentials with hydrogen-bond acidity and basicity scales. We show the newly derived MIFs vary in strength for various ring substitution in accordance with chemical intuition. Finally, we report that this derived parameter, when extended to non-HB atoms, can also be used to estimate sites of reaction.

| S-EPMC9291213 | biostudies-literature

Chemistry-informed molecular graph as reaction descriptor for machine-learned retrosynthesis planning.

Project description:Infusing "chemical wisdom" should improve the data-driven approaches that rely exclusively on historical synthetic data for automatic retrosynthesis planning. For this purpose, we designed a chemistry-informed molecular graph (CIMG) to describe chemical reactions. A collection of key information that is most relevant to chemical reactions is integrated in CIMG:NMR chemical shifts as vertex features, bond dissociation energies as edge features, and solvent/catalyst information as global features. For any given compound as a target, a product CIMG is generated and exploited by a graph neural network (GNN) model to choose reaction template(s) leading to this product. A reactant CIMG is then inferred and used in two GNN models to select appropriate catalyst and solvent, respectively. Finally, a fourth GNN model compares the two CIMG descriptors to check the plausibility of the proposed reaction. A reaction vector is obtained for every molecule in training these models. The chemical wisdom of reaction propensity contained in the pretrained reaction vectors is exploited to autocategorize molecules/reactions and to accelerate Monte Carlo tree search (MCTS) for multistep retrosynthesis planning. Full synthetic routes with recommended catalysts/solvents are predicted efficiently using this CIMG-based approach.

| S-EPMC9564830 | biostudies-literature

Quantum mechanical/molecular mechanical free energy simulations of the self-cleavage reaction in the hepatitis delta virus ribozyme.

Project description:The hepatitis delta virus (HDV) ribozyme catalyzes a self-cleavage reaction using a combination of nucleobase and metal ion catalysis. Both divalent and monovalent ions can catalyze this reaction, although the rate is slower with monovalent ions alone. Herein, we use quantum mechanical/molecular mechanical (QM/MM) free energy simulations to investigate the mechanism of this ribozyme and to elucidate the roles of the catalytic metal ion. With Mg(2+) at the catalytic site, the self-cleavage mechanism is observed to be concerted with a phosphorane-like transition state and a free energy barrier of ∼13 kcal/mol, consistent with free energy barrier values extrapolated from experimental studies. With Na(+) at the catalytic site, the mechanism is observed to be sequential, passing through a phosphorane intermediate, with free energy barriers of 2-4 kcal/mol for both steps; moreover, proton transfer from the exocyclic amine of protonated C75 to the nonbridging oxygen of the scissile phosphate occurs to stabilize the phosphorane intermediate in the sequential mechanism. To explain the slower rate observed experimentally with monovalent ions, we hypothesize that the activation of the O2' nucleophile by deprotonation and orientation is less favorable with Na(+) ions than with Mg(2+) ions. To explore this hypothesis, we experimentally measure the pKa of O2' by kinetic and NMR methods and find it to be lower in the presence of divalent ions rather than only monovalent ions. The combined theoretical and experimental results indicate that the catalytic Mg(2+) ion may play three key roles: assisting in the activation of the O2' nucleophile, acidifying the general acid C75, and stabilizing the nonbridging oxygen to prevent proton transfer to it.

| S-EPMC3954522 | biostudies-literature

Fundamental reaction pathway for peptide metabolism by proteasome: insights from first-principles quantum mechanical/molecular mechanical free energy calculations.

Project description:Proteasome is the major component of the crucial non-lysosomal protein degradation pathway in the cells, but the detailed reaction pathway is unclear. In this study, first-principles quantum mechanical/molecular mechanical free energy calculations have been performed to explore, for the first time, possible reaction pathways for proteasomal proteolysis/hydrolysis of a representative peptide, succinyl-leucyl-leucyl-valyl-tyrosyl-7-amino-4-methylcoumarin (Suc-LLVY-AMC). The computational results reveal that the most favorable reaction pathway consists of six steps. The first is a water-assisted proton transfer within proteasome, activating Thr1-O(?). The second is a nucleophilic attack on the carbonyl carbon of a Tyr residue of substrate by the negatively charged Thr1-O(?), followed by the dissociation of the amine AMC (third step). The fourth step is a nucleophilic attack on the carbonyl carbon of the Tyr residue of substrate by a water molecule, accompanied by a proton transfer from the water molecule to Thr1-N(z). Then, Suc-LLVY is dissociated (fifth step), and Thr1 is regenerated via a direct proton transfer from Thr1-N(z) to Thr1-O(?). According to the calculated energetic results, the overall reaction energy barrier of the proteasomal hydrolysis is associated with the transition state (TS3(b)) for the third step involving a water-assisted proton transfer. The determined most favorable reaction pathway and the rate-determining step have provided a reasonable interpretation of the reported experimental observations concerning the substituent and isotopic effects on the kinetics. The calculated overall free energy barrier of 18.2 kcal/mol is close to the experimentally derived activation free energy of ?18.3-19.4 kcal/mol, suggesting that the computational results are reasonable.

| S-EPMC3851579 | biostudies-literature

Quantum mechanical reaction rate constants by vibrational configuration interaction: the OH + H2->H2O + H reaction as a function of temperature.

Project description:The thermal rate constant of the 3D OH + H(2)-->H(2)O + H reaction was computed by using the flux autocorrelation function, with a time-independent square-integrable basis set. Two modes that actively participate in bond making and bond breaking were treated by using 2D distributed Gaussian functions, and the remaining (nonreactive) modes were treated by using harmonic oscillator functions. The finite-basis eigenvalues and eigenvectors of the Hamiltonian were obtained by solving the resulting generalized eigenvalue equation, and the flux autocorrelation function for a dividing surface optimized in reduced-dimensionality calculations was represented in the basis formed by the eigenvectors of the Hamiltonian. The rate constant was obtained by integrating the flux autocorrelation function. The choice of the final time to which the integration is carried was determined by a plateau criterion. The potential energy surface was from Wu, Schatz, Lendvay, Fang, and Harding (WSLFH). We also studied the collinear H + H(2) reaction by using the Liu-Siegbahn-Truhlar-Horowitz (LSTH) potential energy surface. The calculated thermal rate constant results were compared with reported values on the same surfaces. The success of these calculations demonstrates that time-independent vibrational configuration interaction can be a very convenient way to calculate converged quantum mechanical rate constants, and it opens the possibility of calculating converged rate constants for much larger reactions than have been treated until now.

| S-EPMC1100741 | biostudies-literature

Quantum mechanical investigation of the effect of catalyst fluorination in the intermolecular asymmetric Stetter reaction.

Project description:The asymmetric intermolecular Stetter reaction was investigated using the B3LYP and M06-2X functionals. Fluorination of a triazolium bicyclic catalyst had been found to significantly influence reaction yields and enantiomeric ratios. Computations indicate that the improved reactivity of the fluorinated catalyst is due to better electrostatic interactions between the nitroalkene and catalyst. Computational investigations of preferred conformations of the ground state catalyst and acyl anion equivalent, and the transition structures leading to both enantiomers of the products, are reported.

| S-EPMC3143204 | biostudies-literature

Machine Learning Descriptors for Data-Driven Catalysis Study.

Project description:Traditional trial-and-error experiments and theoretical simulations have difficulty optimizing catalytic processes and developing new, better-performing catalysts. Machine learning (ML) provides a promising approach for accelerating catalysis research due to its powerful learning and predictive abilities. The selection of appropriate input features (descriptors) plays a decisive role in improving the predictive accuracy of ML models and uncovering the key factors that influence catalytic activity and selectivity. This review introduces tactics for the utilization and extraction of catalytic descriptors in ML-assisted experimental and theoretical research. In addition to the effectiveness and advantages of various descriptors, their limitations are also discussed. Highlighted are both 1) newly developed spectral descriptors for catalytic performance prediction and 2) a novel research paradigm combining computational and experimental ML models through suitable intermediate descriptors. Current challenges and future perspectives on the application of descriptors and ML techniques to catalysis are also presented.

| S-EPMC10401178 | biostudies-literature

OmicsDI is part of the ELIXIR infrastructure

OmicsDI is an Elixir interoperability service. Learn more ›

Tweets

OmicsDI Databases

PRIDE
PeptideAtlas
MassIVE
JPOST Repository
Physiome Model Repository

EGA
EVA
ENA
LINCS
PAXDB
Cell Collective

MetaboLights
Metabolomics Workbench
MetabolomeExpress
GNPS
BioModels
FAIRDOMHub

ArrayExpress
dbGaP
ExpressionAtlas
GEO
NODE

Information

Databases
Help
API
Contact us
Code on GitHub
Terms of use
Submit Data