ConTIG: Continuous representation learning on temporal interaction graphs.

Representation learning on temporal interaction graphs (TIG) aims to model complex networks with the dynamic evolution of interactions on a wide range of web and social graph applications. However, most existing works on TIG either (a) rely on discretely updated node embeddings merely when an interaction occurs that fail to capture the continuous evolution of embedding trajectories of nodes, or (b) overlook the rich temporal patterns hidden in the ever-changing graph data that presumably lead to sub-optimal models. In this paper, we propose a two-module framework named ConTIG, a novel representation learning method on TIG that captures the continuous dynamic evolution of node embedding trajectories. With two essential modules, our model exploits three-fold factors in dynamic networks including latest interaction, neighbor features, and inherent characteristics. In the first update module, we employ a continuous inference block to learn the nodes' state trajectories from time-adjacent interaction patterns using ordinary differential equations. In the second transform module, we introduce a self-attention mechanism to predict future node embeddings by aggregating historical temporal interaction information. Experiment results demonstrate the superiority of ConTIG on temporal link prediction, temporal node recommendation, and dynamic node classification tasks of four datasets compared with a range of state-of-the-art baselines, especially for long-interval interaction prediction.

Amelioration of Cyclosporine A-Induced Acute Nephrotoxicity by Cordyceps cicadae Mycelia via Mg+2 Reabsorption and the Inhibition of GRP78-IRE1-CHOP Pathway: In Vivo and In Vitro.

Fruiting bodies of Cordyceps cicadae (CC) have been reported to have a therapeutic effect in chronic kidney disease. Due to the rare and expensive resources from natural habitats, artificially cultivated mycelia using submerged liquid cultivation of CC (CCM) have been recently developed as an alternative to scarce sources of CC. However, little is known regarding potential protective effects of CCM against cyclosporine A (CsA)-induced acute nephrotoxicity in vivo and in vitro. In this study, male Sprague-Dawley rats were divided into six groups: control, CCM (40 mg and 400 mg/kg, orally), CsA (10 mg/kg, oral gavage), and CsA + CCM (40 mg and 400 mg/kg, orally). At the end of the study on day 8, all rats were sacrificed, and the blood and kidneys retrieved. CsA-induced acute nephrotoxicity was evident by increased levels of blood urea nitrogen (BUN). Levels of the endoplasmic reticulum (ER) resident chaperone glucose regulated protein 78 (GRP 78) were increased significantly in rats with acute nephrotoxicity. BUN and GRP 78 were significantly ameliorated in synchronous oral groups of CCM (40 or 400 mg/kg) plus CsA. Examination of hematoxylin and eosin stained kidney tissues revealed that the combined treatment of CCM slightly improved vacuolization in renal tubules upon CsA-induced damage. CsA-induced down-regulation of protein expression of magnesium ion channel proteins and transient receptor potential melastatin 6 and 7 were abolished by the combined treatment of CCM. CCM has the potential to protect the kidney against CsA-induced nephrotoxicity by reducing magnesium ion wasting, tubular cell damage, and ER stress demonstrated further by human renal proximal tubular epithelial cell line HK-2. Our results contribute to the in-depth understanding of the role of polysaccharides and nucleobases as the main secondary metabolites of CCM in the defense system of renal functions in CsA-induced acute nephrotoxicity.

TraverseNet: Unifying Space and Time in Message Passing for Traffic Forecasting.

This article aims to unify spatial dependency and temporal dependency in a non-Euclidean space while capturing the inner spatial-temporal dependencies for traffic data. For spatial-temporal attribute entities with topological structure, the space-time is consecutive and unified while each node's current status is influenced by its neighbors' past states over variant periods of each neighbor. Most spatial-temporal neural networks for traffic forecasting study spatial dependency and temporal correlation separately in processing, gravely impaired the spatial-temporal integrity, and ignore the fact that the neighbors' temporal dependency period for a node can be delayed and dynamic. To model this actual condition, we propose TraverseNet, a novel spatial-temporal graph neural network, viewing space and time as an inseparable whole, to mine spatial-temporal graphs while exploiting the evolving spatial-temporal dependencies for each node via message traverse mechanisms. Experiments with ablation and parameter studies have validated the effectiveness of the proposed TraverseNet, and the detailed implementation can be found from https://github.com/nnzhan/TraverseNet.

MoO3 Nanoparticle Coatings on High-Voltage 5 V LiNi0.5Mn1.5O4 Cathode Materials for Improving Lithium-Ion Battery Performance.

To reduce surface contamination and increase battery life, MoO3 nanoparticles were coated with a high-voltage (5 V) LiNi0.5Mn1.5O4 cathode material by in-situ method during the high-temperature annealing process. To avoid charging by more than 5 V, we also developed a system based on anode-limited full-cell with a negative/positive electrode (N/P) ratio of 0.9. The pristine LiNi0.5Mn1.5O4 was initially prepared by high-energy ball-mill with a solid-state reaction, followed by a precipitation reaction with a molybdenum precursor for the MoO3 coating. The typical structural and electrochemical behaviors of the materials were clearly investigated and reported. The results revealed that a sample of 2 wt.% MoO3-coated LiNi0.5Mn1.5O4 electrode exhibited an optimal electrochemical activity, indicating that the MoO3 nanoparticle coating layers considerably enhanced the high-rate charge-discharge profiles and cycle life performance of LiNi0.5Mn1.5O4 with a negligible capacity decay. The 2 wt.% MoO3-coated LiNi0.5Mn1.5O4 electrode could achieve high specific discharge capacities of 131 and 124 mAh g-1 at the rates of 1 and 10 C, respectively. In particular, the 2 wt.% MoO3-coated LiNi0.5Mn1.5O4 electrode retained its specific capacity (87 mAh g-1) of 80.1% after 500 cycles at a rate of 10 C. The Li4Ti5O12/LiNi0.5Mn1.5O4 full cell based on the electrochemical-cell (EL-cell) configuration was successfully assembled and tested, exhibiting excellent cycling retention of 93.4% at a 1 C rate for 100 cycles. The results suggest that the MoO3 nano-coating layer could effectively reduce side reactions at the interface of the LiNi0.5Mn1.5O4 cathode and the electrolyte, thus improving the electrochemical performance of the battery system.

Structural and Spectroscopic Evidence for a Side-on Fe(III)-Superoxo Complex Featuring Discrete O-O Bond Distances.

The O-O bond length is often used as a structural indicator to determine the valence states of bound O2 ligands in biological metal-dioxygen intermediates and related biomimetic complexes. Here, we report very distinct O-O bond lengths found for three crystallographic forms (1.229(4), 1.330(4), 1.387(2) Šat 100 K) of a side-on iron-dioxygen species. Despite their different O-O bond distances, all forms possess the same electronic structure of Fe(III)-O2 •-, as evidenced by their indistinguishable spectroscopic features. Density functional theory and ab initio calculations, which successfully reproduce spectroscopic parameters, predict a flat potential energy surface of an η2-O2 motif binding to the iron center regarding the O-O distance. Therefore, the discrete O-O bond lengths observed likely arise from differential intermolecular interactions in the second coordination sphere. The work suggests that the O-O distance is not a reliable benchmark to unequivocally identify the valence state of O2 ligands for metal-dioxygen species in O2-utilizing metalloproteins and synthetic complexes.

Graphene Nanosheet-Wrapped Mesoporous La0.8Ce0.2Fe0.5Mn0.5O3 Perovskite Oxide Composite for Improved Oxygen Reaction Electro-Kinetics and Li-O2 Battery Application.

A novel design and synthesis methodology is the most important consideration in the development of a superior electrocatalyst for improving the kinetics of oxygen electrode reactions, such as the oxygen reduction reaction (ORR) and the oxygen evolution reaction (OER) in Li-O2 battery application. Herein, we demonstrate a glycine-assisted hydrothermal and probe sonication method for the synthesis of a mesoporous spherical La0.8Ce0.2Fe0.5Mn0.5O3 perovskite particle and embedded graphene nanosheet (LCFM(8255)-gly/GNS) composite and evaluate its bifunctional ORR/OER kinetics in Li-O2 battery application. The physicochemical characterization confirms that the as-formed LCFM(8255)-gly perovskite catalyst has a highly crystalline structure and mesoporous morphology with a large specific surface area. The LCFM(8255)-gly/GNS composite hybrid structure exhibits an improved onset potential and high current density toward ORR/OER in both aqueous and non-aqueous electrolytes. The LCFM(8255)-gly/GNS composite cathode (ca. 8475 mAh g-1) delivers a higher discharge capacity than the La0.5Ce0.5Fe0.5Mn0.5O3-gly/GNS cathode (ca. 5796 mAh g-1) in a Li-O2 battery at a current density of 100 mA g-1. Our results revealed that the composite's high electrochemical activity comes from the synergism of highly abundant oxygen vacancies and redox-active sites due to the Ce and Fe dopant in LaMnO3 and the excellent charge transfer characteristics of the graphene materials. The as-developed cathode catalyst performed appreciable cycle stability up to 55 cycles at a limited capacity of 1000 mAh g-1 based on conventional glass fiber separators.

A Comprehensive Survey on Graph Neural Networks.

Deep learning has revolutionized many machine learning tasks in recent years, ranging from image classification and video processing to speech recognition and natural language understanding. The data in these tasks are typically represented in the Euclidean space. However, there is an increasing number of applications, where data are generated from non-Euclidean domains and are represented as graphs with complex relationships and interdependency between objects. The complexity of graph data has imposed significant challenges on the existing machine learning algorithms. Recently, many studies on extending deep learning approaches for graph data have emerged. In this article, we provide a comprehensive overview of graph neural networks (GNNs) in data mining and machine learning fields. We propose a new taxonomy to divide the state-of-the-art GNNs into four categories, namely, recurrent GNNs, convolutional GNNs, graph autoencoders, and spatial-temporal GNNs. We further discuss the applications of GNNs across various domains and summarize the open-source codes, benchmark data sets, and model evaluation of GNNs. Finally, we propose potential research directions in this rapidly growing field.

The dual roles of a V(III) centre for substrate binding and oxygen atom abstraction; nitrite reduction mediated by a V(III) complex.

A V(iii) complex bearing a tris(thiolato)phoshine derivative mediates the reduction of nitrite without the assistance of external protons or oxophilic substrates. The metal site plays dual roles for nitrite binding and deoxygenation. The reaction is monitored by spectroscopy combined with isotopic labeling experiments. The formed product, a {VNO}4 species, is isolated and characterized.

Statistical modeling and knowledge-based segmentation of cerebral artery based on TOF-MRA and MR-T1.

BACKGROUND AND OBJECTIVE: For cerebrovascular segmentation from time-of-flight (TOF) magnetic resonance angiography (MRA), the focused issues are segmentation accuracy, vascular network coverage ratio, and cerebral artery and vein (CA/CV) separation. Therefore, cerebral artery segmentation is a challenging work, while a complete solution is lacking so far. METHODS: The preprocessing of skull-stripping and Hessian-based feature extraction is first implemented to acquire an indirect prior knowledge of vascular distribution and shape. Then, a novel intensity- and shape-based Markov statistical modeling is proposed for complete cerebrovascular segmentation, where our low-level process employs a Gaussian mixture model to fit the intensity histogram of the skull-stripped TOF-MRA data, while our high-level process employs the vascular shape prior to construct the energy function. To regularize the individual data processes, Markov regularization parameter is automatically estimated by using a machine-learning algorithm. Further, cerebral artery and vein (CA/CV) separation is explored with a series of morphological logic operations, which are based on a direct priori knowledge on the relationship of arteriovenous topology and brain tissues in between TOF-MRA and MR-T1. RESULTS: We employed 109 sets of public datasets from MIDAS for qualitative and quantitative assessment. The Dice similarity coefficient, false negative rate (FNR), and false positive rate (FPR) of 0.933, 0.158, and 0.091% on average, as well as CA/CV separation results with the agreement, FNR, and FPR of 0.976, 0.041, and 0.022 on average. For clinical visual assessment, our methods can segment various sizes of the vessel in different contrast region, especially performs better on vessels of small size in low contrast region. CONCLUSION: Our methods obtained satisfying results in visual and quantitative evaluation. The proposed method is capable of accurate cerebrovascular segmentation and efficient CA/CV separation. Further, it can stimulate valuable clinical applications on the computer-assisted cerebrovascular intervention according to the neurosurgeon's recommendation.

Reactivity of a Fe(III)-Bound Methoxide Supported with a Tris(thiolato)phosphine Ligand: Activation of C-Cl Bond in CH2Cl2 by Nucleophilic Attack of a Fe(III)-OCH3 Moiety.

Two mononuclear nonheme Fe(III) complexes, [PPh4][Fe(III)(PS3″)(OCH3)] (1) and [PPh4][Fe(III)(PS3″)(Cl)] (2), supported by a tris(benzenethiolato)phosphine derivative PS3″ (PS3″ = P(C6H3-3-Me3Si-2-S)3(3-)) have been synthesized and characterized. The structures resolved from X-ray crystallography show that Fe(III) centers in both complexes adopt distorted trigonal-bipyramidal geometry with a methoxide or a chloride binding in the axial position. The magnetic data for both are consistent with intermediate-spin Fe(III) centers with a C3 symmetry (S = 3/2 ground state). The bound methoxide in 1 is labile and can be replaced by a CH3CN molecule. The forming Fe(III)-CH3CN species can be further reduced by cobaltcene quantitatively to a stable Fe(II)-CH3CN complex, [Fe(PS3″)(CH3CN)](-). One-electron oxidation of 2 by ferrocenium gave a Fe(IV) analogue, [Fe(IV)(PS3″)(Cl)]. Importantly, the Fe(III)-OCH3 moiety in complex 1 acts as a strong nucleophile that activates the C-Cl bond in CH2Cl2, leading to the formation of complex 2 quantitatively. Complex 1 also reacts with other electrophiles, benzyl chloride and benzyl bromide, to generate Fe(III)-X species (X = Cl or Br). The reactions were investigated and monitored by UV-vis-NIR, NMR, and ESI-MS spectroscopies.

The optimal extracting process, manufacturing technique and biological evaluation of Lithospermum erythrorhizon microcapsules.

Lithospermum erythrorhizon has been proved to be anti-inflammatory, by recent studies. This study extracts L. erythrorhizon with ethanol at various solid-liquid ratios (1:4, 1:6, 1:8, and 1:12), extraction temperatures (40°C, 50°C, and 60°C), and extraction times (4, 24 and 36h) in order to determine the optimal parameters. The optimal parameters are extracted and condensed into L. erythrorhizon extract; then the antibacterial property and cell compatibility of L. erythrorhizon extract are evaluated with various concentrations of L. erythrorhizon extract solution and different weights of L. erythrorhizon extract powder, respectively. The concentrations of solution are 0.1mg/ml, 0.5mg/ml, 1.0mg/ml, and 2.0mg/ml and ethanol is chosen as the solvent, and different weights of powder are varied as 0.1mg, 1.0mg, 2.0mg, and 10mg. The cell viability test and animal study are performed on L. erythrorhizon microcapsules. The experiment results show that sodium alginate/pectin L. erythrorhizon (SPL) microcapsules possess a 120-hour drug release. The results of cell viability and animal study show that the L. erythrorhizon microcapsules (SPL) have good cell viability (99%) and can help in the wound healing process (the wound size reduction reaches 91.3% on Day 11).

Direct optical switching of bistable cholesteric textures in chiral azobenzene-doped liquid crystals.

The direct optical switching of bistable cholesteric textures (i.e., planar and focal conic textures) in chiral azobenzene-doped liquid crystals (LCs) is demonstrated. Chiral azobenzene is a chiral dopant with optically tuned helical twisting power that results from the photo-isomerization between trans- and cis- isomers via exposure to UV or visible light. The pitch length of the material can be optically and repeatedly elongated and shortened. With regard to free energy, LCs tend to be stable at planar (focal conic) textures when pitch length is elongated (shortened) by exposure to UV (visible) light. Thus, direct optical switchable LC displays are investigated.

Oligonucleotide-functionalized silver nanoparticle extraction and laser-induced fluorescence for ultrasensitive detection of mercury(II) ion.

This study describes the development of a simple, sensitive, and selective detection system for Hg(2+) ion by combining nanoparticle extraction, fluorescent dye labeling, and flow injection analysis (FIA) detection. Repeats of 33 thymine nucleotides-functionalized silver nanoparticles (T(33)-AgNPs) specifically capture Hg(2+) from aqueous solution through the coordination between T(33) and Hg(2+). Meanwhile, Hg(2+) ion drives a T(33) conformational change from a random coil to a folded structure. The T(33)-Hg(2+)complexes adsorbed on the NP surface were collected from the initial sample by centrifugation, and they were then detached from the NP surface by addition of H(2)O(2). The T(33)-Hg(2+) complexes preferentially bind to SYBR Green I (SG), enhancing the SG fluorescence. By contrast, SG fluoresces only weakly in the presence of T(33) alone. The extraction efficiency of Hg(2+) was highly dependent on polythymine length, the concentration of T(33)-AgNPs, and the incubaton time of T(33)-AgNPs with Hg(2+). Under optimal extraction and labeling conditions, FIA detection showed the limit of detection (at a signal-to-noise ratio of three) for Hg(2+)of 3 pM. The selectivity of our analytical system is more than 1000-fold for Hg(2+) over any metal ions. We validated the applicability of this system for the determination of Hg(2+) concentrations in tap water.

Colorimetric assay for S-adenosylhomocysteine hydrolase activity and inhibition using fluorosurfactant-capped gold nanoparticles.

This study reports a simple colorimetric method for the sensitive detection of S-adenosylhomocysteine hydrolase (SAHH) activity and inhibition using fluorosurfactant-capped gold nanoparticles (FSN-AuNPs). FSN stabilizes the AuNPs against conditions of high ionic strength, and FSN-AuNPs are merely aggregated in the presence of homocysteine (HCys) and cysteine. Because of this feature, FSN-AuNPs were found to be dispersed in the presence of S-adenosylhomocysteine (SAH) that lacks a free thiol group. After SAHH catalyzed the hydrolysis of SAH, the produced HCys molecules were bound to the surface of AuNPs through the formation of Au-S bonds. As a result, the nanoparticle (NP) aggregation occurred through electrostatic attraction between each HCys-attached AuNP. This approach had a minimum detectable concentration of 100 units/L (~6 nM). Additionally, because adenosine analogs are capable of inhibiting SAHH activity, the addition of adenosine analogs to a solution containing SAH and SAHH resulted in the suppression of hydrolyzed SAH-induced NP aggregation. Adenosine analogs exhibited the following trend in the half-maximal inhibitory concentrations: adenosine > adenosine monophosphate > adenosine diphosphate ~ adenosine triphosphate. We have demonstrated that the combination of SAHH inhibition and FSN-AuNPs can be utilized for the selective detection of adenosine.

Representing images using points on image surfaces.

This paper presents a new approach to represent an image by "verge points," which are defined as high-curvature points on the image surface. This representation offers a compact and reversible way to preserve the essence of the original image. Various applications, such as compression, edge detection, image enhancement, and image editing, can be achieved based on this representation. In this paper, the whole procedure for verge point representation is presented. Based on these verge points, image reconstruction can be easily achieved via iterative linear interpolation. These extracted verge points with compatible properties are further linked into verge curves to offer more compact representation. Progressive representation is also developed based on a multiscale extraction scheme. Some potential applications are then presented to demonstrate the versatility of this representation.