Using principal component analysis for neural network high-dimensional potential energy surface

A systematical strategy for quality markers screening of different methods processing Platycodonis radix based on phytochemical analysis and the impact on Chronic Obstructive Pulmonary Disease

ETHNOPHARMACOLOGICAL RELEVANCE

Baihezhijiegeng is a processed product of Platycodonis radix, and it's effective in the treatment of Chronic Obstructive Pulmonary Disease (COPD). However, the specific mechanism of action has not been reported in the literature.

AIM OF THE STUDY

We attempted to evaluate the phytochemical composition and pharmaco-dynamics of Platycodon grandiflorum (PG) and BJ to clarify the mechanism behind the expectorant effect of BJ.

MATERIALS AND METHODS

We integrated the ultra-high-performance liquid chromatography-linear trap quadrupole orbitrap velos mass spectrometry (UPLC-LTQ Orbitrap MS/MS) and the ultra-performance liquid chromatography quadrupole-time-of-flight tandem mass spectrometry (UPLC-Q-TOF-MS/MS) methods to identify the chemical constituents of PG and BJ. Moreover, correlation and multivariate statistical analyses were utilized to seek the candidate quality markers of PG and BJ. Analysis of effective herbal chemical components using UPLC-Q-TOF-MS/MS and retrieval of COPD disease targets from OMIM, TTD, GeneCard databases. Protein-protein interaction (PPI) and topology analyses were performed using the String database and Cytoscape 3.7.2 software; gene ontology (GO) functional enrichment analysis and Kyoto encyclopedia of genes and genomes (KEGG) pathway enrichment analysis were performed using the Metescape platform on common targets. Moreover, we used molecular docking to predict the potential mechanism of quality markers for developing anti-COPD activity. Simultaneously, the model of COPD was established by exposing the animals to cigarette smoke combined with a tracheal drip injection of lipopolysaccharide (LPS). Using the ELISA method, quantitative real-time polymerase chain reaction (qRT-PCR) and western blotting (WB) to determine tumor-necrosis factor (TNF)-α, interleukin (IL)-1β, IL-6, and matrix metalloproteinase (MMP)9 levels in serum and IL-4, IL-10, IFN-γ levels, epidermal growth factor receptor (EGFR) and MUC5AC expression in lung tissue of COPD rats to explore the therapeutic effects of PG and BJ on the COPD rat model.

RESULTS

The chemical identification of JG and PG extracts using UPLC-Q-TOF-MS/MS and UPLC-LTQ Orbitrap MS/MS showed 71 compounds, including 47 saponins, 16 phenolic acids, four flavonoids, and four other components. The multivariate statistical analysis showed that seven quality markers were screened. Network pharmacology results showed a role in biological processes such as cellular response to hydrogen peroxide, positive regulation of pri-miRNA transcription from RNA polymerase II promoter, molecular functions such as oxidoreductase activity, acting on NAD(P)H, quinone or similar compound as acceptor, bile acid binding and other molecular functions. In COPD rats, histopathological findings depicted that BJ administration could effectively inhibit inflammatory cell infiltration and mucus hypersecretion, and improve the lung pathological status in rats with COPD. Moreover, BJ could significantly decrease TNF-α, IL-1β, IL-6, and matrix metalloproteinase (MMP)9 levels in the serum and interferon (IFN)-γ levels in lung tissues of rats with COPD (p < 0.01), and significantly increase IL-4 and IL-10 levels in their lung tissues (p < 0.01), suggesting its inhibition of the inflammatory response in vivo. Additionally, EGFR and MUC5AC were reduced in the lung tissues of rats with COPD and airway mucus hypersecretion in rats with COPD.

CONCLUSION

This study revealed the material basis of PG and BJ for anti-COPD activity and discovered the quality markers of PG and BJ which could affect the anti-COPD activity. The therapeutic effects of BJ may be attributed to the regulation of the inflammatory mediators and mediation of the EGFR/MUC5AC pathway in rats with COPD.

How to train a neural network potential

The introduction of modern Machine Learning Potentials (MLPs) has led to a paradigm change in the development of potential energy surfaces for atomistic simulations. By providing efficient access to energies and forces, they allow us to perform large-scale simulations of extended systems, which are not directly accessible by demanding first-principles methods. In these simulations, MLPs can reach the accuracy of electronic structure calculations, provided that they have been properly trained and validated using a suitable set of reference data. Due to their highly flexible functional form, the construction of MLPs has to be done with great care. In this Tutorial, we describe the necessary key steps for training reliable MLPs, from data generation via training to final validation. The procedure, which is illustrated for the example of a high-dimensional neural network potential, is general and applicable to many types of MLPs.

Machine Learning Frontier Orbital Energies of Nanodiamonds

Nanodiamonds have a wide range of applications including catalysis, sensing, tribology, and biomedicine. To leverage nanodiamond design via machine learning, we introduce the new data set ND5k, consisting of 5089 diamondoid and nanodiamond structures and their frontier orbital energies. ND5k structures are optimized via tight-binding density functional theory (DFTB) and their frontier orbital energies are computed using density functional theory (DFT) with the PBE0 hybrid functional. From this data set we derive a qualitative design suggestion for nanodiamonds in photocatalysis. We also compare recent machine learning models for predicting frontier orbital energies for similar structures as they have been trained on (interpolation on ND5k), and we test their abilities to extrapolate predictions to larger structures. For both the interpolation and extrapolation task, we find the best performance using the equivariant message passing neural network PaiNN. The second best results are achieved with a message passing neural network using a tailored set of atomic descriptors proposed here.

Artificial Neural Network-Derived Unified Six-Dimensional Potential Energy Surface for Tetra Atomic Isomers of the Biogenic [H, C, N, O] System

Recognition of different structural patterns in different potential energy surface regions, such as in isomerizing quasilinear tetra atomic molecules, is important for understanding the details of underlying physics and chemistry. In this respect, using three variants of artificial neural networks (ANNs), we investigated the six-dimensional (6-D) singlet potential energy surfaces (PES) of tetra atomic isomers of the biogenic [H, C, N, O] system. At first, we constructed a separate ANN potential for each of the studied isomers. In the next step, a comparative assessment of the separate ANN models led to the setting up of a unified 6-D singlet PES equally and accurately describing all studied isomers. The constructed unified model yields relative energies comparable to those obtained either from the gold standard CCSD(T) method or from separate ANNs for each of the studied isomers. The accuracy of the unified singlet PES is on the order of 10^-4 Hartrees (0.1 kcal/mol). The developed PES in this work captures the main features of nonlinear and quasilinear tetra atomic isomers of this biogenic system.

Reaction Space Projector (ReSPer) for Visualizing Dynamic Reaction Routes Based on Reduced-Dimension Space

To analyze chemical reaction dynamics based on a reaction path network, we have developed the "Reaction Space Projector" (ReSPer) method with the aid of the dimensionality reduction method. This program has two functions: the construction of a reduced-dimensionality reaction space from a molecular structure dataset, and the projection of dynamic trajectories into the low-dimensional reaction space. In this paper, we apply ReSPer to isomerization and bifurcation reactions of the Au₅ cluster and succeed in analyzing dynamic reaction routes involved in multiple elementary reaction processes, constructing complicated networks (called "closed islands") of nuclear permutation-inversion (NPI) isomerization reactions, and elucidating dynamic behaviors in bifurcation reactions with reference to bundles of trajectories. Interestingly, in the second application, we find a correspondence between the contribution ratios in the ability to visualize and the symmetry of the morphology of closed islands. In addition, the third application suggests the existence of boundaries that determine the selectivity in bifurcation reactions, which was discussed in the phase space. The ReSPer program is a versatile and robust tool to clarify dynamic reaction mechanisms based on the reduced-dimensionality reaction space without prior knowledge of target reactions.

Artificial Neural Networks for Pyrolysis, Thermal Analysis, and Thermokinetic Studies: The Status Quo

Artificial neural networks (ANNs) are a method of machine learning (ML) that is now widely used in physics, chemistry, and material science. ANN can learn from data to identify nonlinear trends and give accurate predictions. ML methods, and ANNs in particular, have already demonstrated their worth in solving various chemical engineering problems, but applications in pyrolysis, thermal analysis, and, especially, thermokinetic studies are still in an initiatory stage. The present article gives a critical overview and summary of the available literature on applying ANNs in the field of pyrolysis, thermal analysis, and thermokinetic studies. More than 100 papers from these research areas are surveyed. Some approaches from the broad field of chemical engineering are discussed as the venues for possible transfer to the field of pyrolysis and thermal analysis studies in general. It is stressed that the current thermokinetic applications of ANNs are yet to evolve significantly to reach the capabilities of the existing isoconversional and model-fitting methods.

Hybrid localized graph kernel for machine learning energy-related properties of molecules and solids

Nowadays, the coupling of electronic structure and machine learning techniques serves as a powerful tool to predict chemical and physical properties of a broad range of systems. With the aim of improving the accuracy of predictions, a large number of representations for molecules and solids for machine learning applications has been developed. In this work we propose a novel descriptor based on the notion of molecular graph. While graphs are largely employed in classification problems in cheminformatics or bioinformatics, they are not often used in regression problem, especially of energy-related properties. Our method is based on a local decomposition of atomic environments and on the hybridization of two kernel functions: a graph kernel contribution that describes the chemical pattern and a Coulomb label contribution that encodes finer details of the local geometry. The accuracy of this new kernel method in energy predictions of molecular and condensed phase systems is demonstrated by considering the popular QM7 and BA10 datasets. These examples show that the hybrid localized graph kernel outperforms traditional approaches such as, for example, the smooth overlap of atomic positions and the Coulomb matrices.

Machine learning meets chemical physics

Over recent years, the use of statistical learning techniques applied to chemical problems has gained substantial momentum. This is particularly apparent in the realm of physical chemistry, where the balance between empiricism and physics-based theory has traditionally been rather in favor of the latter. In this guest Editorial for the special topic issue on "Machine Learning Meets Chemical Physics," a brief rationale is provided, followed by an overview of the topics covered. We conclude by making some general remarks.