Design high-entropy carbide ceramics from machine learning

High-entropy ceramics (HECs) have shown great application potential under demanding conditions, such as high stresses and temperatures. However, the immense phase space poses great challenges for the rational design of new high-performance HECs. In this work, we develop machine-learning (ML) models to discover high-entropy ceramic carbides (HECCs). Built upon attributes of HECCs and their constituent precursors, our ML models demonstrate a high prediction accuracy (0.982). Using the well-trained ML models, we evaluate the single-phase probability of 90 HECCs that are not experimentally reported so far. Several of these predictions are validated by our experiments. We further establish the phase diagrams for non-equiatomic HECCs spanning the whole composition space by which the single-phase regime can be easily identified. Our ML models can predict both equiatomic and non-equiatomic HECs based solely on the chemical descriptors of constituent transition-metal-carbide precursors, which paves the way for the high-throughput design of HECCs with superior properties.

Predicting Ruthenium Catalysed Hydrogenation of Esters using Machine Learning

Catalytic hydrogenation of esters is a sustainable approach for the production of fine chemicals, and pharmaceutical drugs. However, the efficiency and cost of catalysts are often the bottlenecks in the commercialization of such technologies. The conventional approach of catalyst discovery is based on empiricism that makes the discovery process time-consuming and expensive. There is an urgent need to develop effective approaches to discover efficient catalysts for hydrogenation reactions. We demonstrate here the approach of machine learning for the prediction of out-comes for the catalytic hydrogenation of esters. Our models can predict the reaction yields with high mean accuracies of up to 91% (test set) and suggest that the use of certain chemical descriptors selectively can result in a more accurate model. Furthermore, cata-lysts and some of their corresponding descriptors can also be pre-dicted with mean accuracies of 85%, and >90%, respectively.

Comparative study of DFT, AM1, and PM3 optimization methods modeling new psychotropic amphetamines

In order to improve the determination of physico-chemical descriptors of organic substances, it is necessary to optimize the computational representations of their molecular structures. The computerized estimation of these descriptors is especially useful in the case of drugs of abuse, as clinical studies are not recommended for new compounds belonging to classes known to have a high toxicity. The optimization of molecular structures allows the improvement of the accuracy in determining the physico-chemical descriptors and consequently of the associated quantitative structure - property or structure - activity relationships (QSAR). We are presenting a comparative study regarding the effect of the AM1, PM3, and DFT optimization methods, which were applied in the case of a series of new psychotropic amphetamines for their physico-chemical characterization based on molecular descriptors.

The Inhibitory Effect of Some Natural Bioactive D-Glucopyranoside Derivatives against SARS-CoV-2 Main Protease (Mpro) and Spike Protease (Spro)

Outbreak of coronavirus seems to have exacerbated across the globe, but drugs have not been discovered till now. Due to having the antiviral activity of D-glucopyranoside derivatives, this study was designed to examine as the inhibitor by in sillico study against the main protease (Mpro) and Spike protease (Spro) of SARS-CoV-2. First, these derivatives were optimised by Density Functional Theory (DFT). The observation of this study was monitored by molecular docking tools calculating the binding affinities. Afterwards, the ligand interaction with protein was accounted for selecting the how to bind of active sites of the protein. Next, the root means square deviation (RMSD) and root mean square fluctuation (RMSF) were illustrated for determining the stability of the docked complex. Finally, AMDET properties were calculated as well as the Lipisinki rule. All of the derivates showed a binding affinity more than -6.0 kcal/mol while derivatives 2, 3, and 9 were the best-bonded scoring inhibitor against Mpro and Spro. In addition, the chemical descriptors were more supportive tools as an inhibitor, and the Lipisinki rule was satisfied for maximum molecules as a drug. Besides, D-glucopyranoside derivatives may be predicted that they are non-carcinogenic and low toxic for both aquatic and non-aquatic species.

Computational Intelligence Model of Orally Disintegrating Tablets: An Attempt to Explain Disintegration Process

We obtained a curated database based on the database published elsewhere. Chemical descriptors were introduced as characteristics of active pharmaceutical ingredients (APIs). We used H2O AutoML platform in order to develop a Deep Learning model and SHAP method to explain its predictions. Obtained results were satisfactory with NRMSE of 8.1% and R2 of 0.84. Finally, we identified critical parameters affecting the process of disintegration of directly compressed ODTs.

HOMO-LUMO photosensitization analyses of coronene-cytosine complexes

Photosensitization analyses of models of (–HC = CH–)n assisted coronene-cytosine complexes assigned by Cor-n-Cyt; n varying by 0, 1, 2, and 3, were investigated in this work by performing density functional theory (DFT) calculations. The investigated models were optimized and chemical descriptors were evaluated. To achieve the goal of this work, energy levels of the highest occupied and the lowest unoccupied molecular orbitals (HOMO and LUMO) were evaluated to reach the absorption energy requirement for innovating photosensitizer (PS) compounds. The models indicated that the complex formations could help the structures to participate in interactions easier than the singular models, in which HOMO-LUMO descriptors indicated lower required absorption energy for them to increase their safety for human health level. The required absorption energies of complexes with n = 0, 1, and 2, were in ultraviolet (UV) region whereas that of complex with n = 3 was moved to visible region. In this regard, the idea of new PS compounds innovation was examined here to introduce Cor-n-Cyt complexes for possible applications in photodynamic therapy (PDT).