Learning size-adaptive molecular substructures for explainable drug–drug interaction prediction by substructure-aware graph neural network

Drug–drug interactions (DDIs) can trigger unexpected pharmacological effects on the body, and the causal mechanisms are often unknown. Graph neural networks (GNNs) have been developed to better understand DDIs. However, identifying key substructures that contribute most to the DDI prediction is a challenge for GNNs. In this study, we presented a substructure-aware graph neural network, a message passing neural network equipped with a novel substructure attention mechanism and a substructure–substructure interaction module (SSIM) for DDI prediction (SA-DDI). Specifically, the substructure attention was designed to capture size- and shape-adaptive substructures based on the chemical intuition that the sizes and shapes are often irregular for functional groups in molecules. DDIs are fundamentally caused by chemical substructure interactions. Thus, the SSIM was used to model the substructure–substructure interactions by highlighting important substructures while de-emphasizing the minor ones for DDI prediction. We evaluated our approach in two real-world datasets and compared the proposed method with the state-of-the-art DDI prediction models. The SA-DDI surpassed other approaches on the two datasets. Moreover, the visual interpretation results showed that the SA-DDI was sensitive to the structure information of drugs and was able to detect the key substructures for DDIs. These advantages demonstrated that the proposed method improved the generalization and interpretation capability of DDI prediction modeling.

SA-DDI is designed to learn size-adaptive molecular substructures for drug–drug interaction prediction and can provide explanations that are consistent with pharmacologists.     

A Novel Method for Pattern Recognition of GIS Partial Discharge via Multi-Information Ensemble Learning

Partial discharge (PD) is the main feature that effectively reflects the internal insulation defects of gas-insulated switchgear (GIS). It is of great significance to diagnose the types of insulation faults by recognizing PD to ensure the normal operation of GIS. However, the traditional diagnosis method based on single feature information analysis has a low recognition accuracy of PD, and there are great differences in the diagnosis effect of various insulation defects. To make the most of the rich insulation state information contained in PD, we propose a novel multi-information ensemble learning for PD pattern recognition. First, the ultra-high frequency and ultrasonic data of PD under four typical defects of GIS are obtained through experiment. Then the deep residual convolution neural network is used to automatically extract discriminative features. Finally, multi-information ensemble learning is used to classify PD types at the decision level, which can complement the shortcomings of the independent recognition of the two types of feature information and has higher accuracy and reliability. Experiments show that the accuracy of the proposed method can reach 97.500%, which greatly improves the diagnosis accuracy of various insulation defects.     

HIGH-FREQUENCY VIBRATIONS OF CORRUGATED CYLINDRICAL PIEZOELECTRIC SHELLS

Coupled extensional and flexural cylindrical vibrations of a corrugated cylindrical piezoelectric shell consisting of multiple pieces of circular cylindrical surfaces smoothly connected along their generatrix are studied. To validate the results for the case of relatively thick shells or equivalently high-frequency modes with short wavelengths, existing analysis is extended by considering shear deformation and rotatory inertia. An analytical solution is obtained. Based on the solution, resonant frequencies and mode shapes are calculated.     

Numerical Study on Entropy Generation of the Multi-Stage Centrifugal Pump

The energy loss of the multi-stage centrifugal pump was investigated by numerical analysis using the entropy generation method with the RNG k-ε turbulence model. Entropy generation due to time-averaged motion and velocity fluctuation was mainly considered. It was found that the entropy generation of guide vanes and impellers account for 71.2% and 23.3% of the total entropy generation under the designed flow condition. The guide vanes are the main hydraulic loss domains and their entropy generation is about 9 W/K, followed by impellers. There are vortices at the tongue of the guide vane inlet as well as flow separations in the impellers, which lead to entropy generation. The fluid impacts the outer surface of the guide vanes, resulting in the increase in entropy generation. There are refluxes near the guide vane tongues which also increase the entropy generation of this part. The entropy generation distribution of the guide vanes and impellers was investigated, which found that the positive guide vane has more entropy generation compared with the reverse guide. The entropy generation of the blade suction surface is higher compared with the pressure surface. This study indicated that the entropy generation method has distinct advantages in the assessment of hydraulic loss.     

Joint Detection of Community and Structural Hole Spanner of Networks in Hyperbolic Space

Community detection and structural hole spanner (the node bridging different communities) identification, revealing the mesoscopic and microscopic structural properties of complex networks, have drawn much attention in recent years. As the determinant of mesoscopic structure, communities and structural hole spanners discover the clustering and hierarchy of networks, which has a key impact on transmission phenomena such as epidemic transmission, information diffusion, etc. However, most existing studies address the two tasks independently, which ignores the structural correlation between mesoscale and microscale and suffers from high computational costs. In this article, we propose an algorithm for simultaneously detecting communities and structural hole spanners via hyperbolic embedding (SDHE). Specifically, we first embed networks into a hyperbolic plane, in which, the angular distribution of the nodes reveals community structures of the embedded network. Then, we analyze the critical gap to detect communities and the angular region where structural hole spanners may exist. Finally, we identify structural hole spanners via two-step connectivity. Experimental results on synthetic networks and real networks demonstrate the effectiveness of our proposed algorithm compared with several state-of-the-art methods.     

Impact of the Acetaldehyde-Mediated Condensation on the Phenolic Composition and Antioxidant Activity of Vitis vinifera L. Cv. Merlot Wine

Acetaldehyde is a critical reactant on modifying the phenolic profile during red wine aging, suggesting that the acetaldehyde-mediated condensation can be responsible for the variation of antioxidant activity during the aging of this beverage. The present study employs exogenous acetaldehyde at six levels of treatment (7.86 ± 0.10–259.02 ± 4.95 mg/L) before the bottle aging of Merlot wines to encourage phenolic modification. Acetaldehyde and antioxidant activity of wine were evaluated at 0, 15, 30, 45, 60 and 75 days of storage, while monomeric and polymeric phenolics were analyzed at 0, 30 and 75 days of storage. The loss of acetaldehyde was fitted to a first-order reaction model, the rate constant (k) demonstrated that different chemical reaction happened in wines containing a different initial acetaldehyde. The disappearance of monomeric phenolics and the formation of polymeric phenolics induced by acetaldehyde could be divided into two phases, the antioxidant activity of wine did not alter significantly in the first phase, although most monomeric phenolics vanished, but the second phase would dramatically reduce the antioxidant activity of wine. Furthermore, a higher level of acetaldehyde could shorten the reaction time of the first phase. These results indicate that careful vinification handling aiming at controlling the acetaldehyde allows one to maintain prolonged biological activity during wine aging.     

文学：概念建构与娱乐主题的沉浮

现代文学概念并非与生俱来,而是现代性理论运作的产物。现代文学概念的定型至少要完成四个方面的理论建构：不同知识体系对接带来的理论扩展,经世致用思想传统的承传,文学语言问题的再认识,叙事文类进入文学视野的中心。相对于现代文学概念的标准涵义,娱乐是一个倍受压抑的主题。尽管遭到了启蒙与革命两种文学观念的强大抵制,但是,作为审美愉悦的一个组成部分,娱乐主题顽强地存留于大众之中。20世纪80年代之后,娱乐主题在“文化产业”的名义下声势浩大地回归,并且获得经济效益参与的评价体系大力支持。对于现代文学概念说来,娱乐主题的重现意味了多方面的理论挑战：娱乐主题的位置,经典与“雅/俗”,大众的构成,商业、市场、传媒体系与文学。这些问题相互联系,持续发酵。现代文学概念应当开启多维的对话与争论,以主动的姿态回应现代社会制造的种种历史境遇。     

An all-organic aqueous potassium dual-ion battery

Benefiting from the environmental friendliness of organic electrodes and the high security of aqueous electrolyte,an all-organic aqueous potassium dual-ion full...     

Intrinsic properties of active sites for hydrogen production from alcohols without coke formation

The detailed reaction pathway and coke formation mechanism over Pt/metal oxide nanoparticles during the steam reforming of ethanol（SRE） at 300℃were studied.The catalysts were prepared by incipient wetness impregnation method and were characterized with CO pulse chemisorption, BET surface measurement,oxygen adsorption,ethanol-TPD,NH₃-TPD,and TPO.The SRE activity of the catalysts with steam/ethanol molar ratio of 3/1 was tested using a continuous fixed-bed reactor.Strong interaction between Pt and supports causes lower H₂ production temperatures and no C₂H₄ formation,while weak interaction leads to C₂H₄ formation and strong bonded CO on Pt particles during ethanol-TPD. H₂ production over Pt-based catalysts is mainly resulted from the decomposition and dehydrogenation of ethanol,and decarbonylation of acetaldehyde.Meanwhile,coke can be formed from acetaldehyde,acetone,C₂H₄ and CO.However,when the interaction between Pt and supports is weak,more coke is formed especially from acetone,C₂H₄ and CO.When the interaction is strong,no coke formation is observed due to high oxygen storage capacity of the catalyst.