Prediction of coronary artery lesions in children with Kawasaki syndrome based on machine learning.

OBJECTIVE: Kawasaki syndrome (KS) is an acute vasculitis that affects children < 5 years of age and leads to coronary artery lesions (CAL) in about 20-25% of untreated cases. Machine learning (ML) is a branch of artificial intelligence (AI) that integrates complex data sets on a large scale and uses huge data to predict future events. The purpose of the present study was to use ML to present the model for early risk assessment of CAL in children with KS by different algorithms. METHODS: A total of 158 children were enrolled from Women and Children's Hospital, Qingdao University, and divided into 70-30% as the training sets and the test sets for modeling and validation studies. There are several classifiers are constructed for models including the random forest (RF), the logistic regression (LR), and the eXtreme Gradient Boosting (XGBoost). Data preprocessing is analyzed before applying the classifiers to modeling. To avoid the problem of overfitting, the 5-fold cross validation method was used throughout all the data. RESULTS: The area under the curve (AUC) of the RF model was 0.925 according to the validation of the test set. The average accuracy was 0.930 (95% CI, 0.905 to 0.956). The AUC of the LG model was 0.888 and the average accuracy was 0.893 (95% CI, 0,837 to 0.950). The AUC of the XGBoost model was 0.879 and the average accuracy was 0.935 (95% CI, 0.891 to 0.980). CONCLUSION: The RF algorithm was used in the present study to construct a prediction model for CAL effectively, with an accuracy of 0.930 and AUC of 0.925. The novel model established by ML may help guide clinicians in the initial decision to make a more aggressive initial anti-inflammatory therapy. Due to the limitations of external validation and regional population characteristics, additional research is required to initiate a further application in the clinic.

Peer effects on rural children's depression: do online games matter?

With the increasing prevalence of depression among children and adolescents, understanding the role of peer contagion in the spread of emotional distress is a critical area of research. The aim of this study is to examine the effect of classmates' depression on a rural child's own depression in China (aged 9-17, N = 1777). The study controls for possible endogeneity of peer effects through the instrumental variable method (the Wald F statistic is significant at the 1% level) and random class assignment data (all students and teachers are randomly assigned to classes).The results indicate that when the average depression score of a rural child's classmates increases by 1 point, that child's own depression score is likely to increase by 0.345 points (p value < .01).This study further finds that the contagion of depression among classmates is more severe for girls and children who play online games, and less severe for children who are cheerful and good-humored. Online games may be an important mechanism through which peer effects operate. Children's classmates' negative emotions are found to increase the occurrence of the children's internalizing behavior in online games, in turn increasing the occurrence of negative emotions in these children themselves.

Understanding the Mechanism of Increased IgG/IgE Reactivity but Decreased Immunodetection Recovery in Thermally Induced Shrimp (Litopenaeus vannamei) Tropomyosin via Multispectroscopic and Molecular Dynamics Simulation Techniques.

Despite the fact that tropomyosin (TM) has highly stable structural characteristics, thermal processing can adversely influence its immunodetection, and the mechanism involved has not been elucidated. Purified TM was heated at various temperatures, and then the IgG/IgE-binding capacity and immunodetection recovery were determined; meanwhile, the structural alterations were analyzed via spectroscopic and molecular dynamics simulation techniques. The obtained results demonstrated that heat-treated TM showed significantly increased IgG/IgE reactivity, confirmed by indirect ELISA and immunoblotting analysis, which might be attributed to the increased structural flexibility, and thus allowed TM to be recognized IgG/IgE easily. However, these structural alterations during thermal processing would contribute to the masking of some epitopes located in TM's surface due to the presence of curled or folded conformation with a considerable reduction of the solvent-accessible surface and radius of gyration, which primarily caused immunodetection recovery reduction in the sandwich ELISA (sELISA) test. The number of antigen binding sites might play a crucial role in a sandwich immunodetection system for sensitive and precise analysis in processed foods.

Insight into the molecular-level details of αs1 casein interactions with IgG: Combining with LC-MS/MS and molecular modelling techniques.

αs1-Casein (αs1-CN) is a major cow milk allergen, while the tertiary structure of αs1-CN and conformational epitopes of αs1-CN have not been clarified. Here, a reasonable three-dimensional structure of αs1-CN was established using ab initio methods, and hot-spot residues and epitopes were investigated by combining molecular dynamics simulation, peptides synthesis, and ELISA. Obtained results demonstrated that the binding mechanism between αs1-CN and IgG was located on three main regions: a helical structure zone (E77-Q97), the flexible loop zone (Y154-T174), and a flexible C-terminal (N190-L198), mainly connecting via hydrogen bond and ionic bonds. The hydrolysates produced by papain with lowest antigenicity (12.43%), which could considerably destroy the essential epitopes of αs1-CN confirmed by epitope synthesis, and LC-MS/MS. The results reported herein would provide novel insights into the interface interactions between αs1-CN and IgG, and prove valuable for developing hypoallergenic infant-formula and peptide vaccines for allergen-specific immunotherapy.

Sentiment Classification of Chinese Tourism Reviews Based on ERNIE-Gram+GCN.

Nowadays, tourists increasingly prefer to check the reviews of attractions before traveling to decide whether to visit them or not. To respond to the change in the way tourists choose attractions, it is important to classify the reviews of attractions with high precision. In addition, more and more tourists like to use emojis to express their satisfaction or dissatisfaction with the attractions. In this paper, we built a dataset for Chinese attraction evaluation incorporating emojis (CAEIE) and proposed an explicitly n-gram masking method to enhance the integration of coarse-grained information into a pre-training (ERNIE-Gram) and Text Graph Convolutional Network (textGCN) (E2G) model to classify the dataset with a high accuracy. The E2G preprocesses the text and feeds it to ERNIE-Gram and TextGCN. ERNIE-Gram was trained using its unique mask mechanism to obtain the final probabilities. TextGCN used the dataset to construct heterogeneous graphs with comment text and words, which were trained to obtain a representation of the document output category probabilities. The two probabilities were calculated to obtain the final results. To demonstrate the validity of the E2G model, this paper was compared with advanced models. After experiments, it was shown that E2G had a good classification effect on the CAEIE dataset, and the accuracy of classification was up to 97.37%. Furthermore, the accuracy of E2G was 1.37% and 1.35% ahead of ERNIE-Gram and TextGCN, respectively. In addition, two sets of comparison experiments were conducted to verify the performance of TextGCN and TextGAT on the CAEIE dataset. The final results showed that ERNIE and ERNIE-Gram combined TextGCN and TextGAT, respectively, and TextGCN performed 1.6% and 2.15% ahead. This paper compared the effects of eight activation functions on the second layer of the TextGCN and the activation-function-rectified linear unit 6 (RELU6) with the best results based on experiments.

ResSUMO: A Deep Learning Architecture Based on Residual Structure for Prediction of Lysine SUMOylation Sites.

Lysine SUMOylation plays an essential role in various biological functions. Several approaches integrating various algorithms have been developed for predicting SUMOylation sites based on a limited dataset. Recently, the number of identified SUMOylation sites has significantly increased due to investigation at the proteomics scale. We collected modification data and found the reported approaches had poor performance using our collected data. Therefore, it is essential to explore the characteristics of this modification and construct prediction models with improved performance based on an enlarged dataset. In this study, we constructed and compared 16 classifiers by integrating four different algorithms and four encoding features selected from 11 sequence-based or physicochemical features. We found that the convolution neural network (CNN) model integrated with residue structure, dubbed ResSUMO, performed favorably when compared with the traditional machine learning and CNN models in both cross-validation and independent tests. The area under the receiver operating characteristic (ROC) curve for ResSUMO was around 0.80, superior to that of the reported predictors. We also found that increasing the depth of neural networks in the CNN models did not improve prediction performance due to the degradation problem, but the residual structure could be included to optimize the neural networks and improve performance. This indicates that residual neural networks have the potential to be broadly applied in the prediction of other types of modification sites with great effectiveness and robustness. Furthermore, the online ResSUMO service is freely accessible.

SWTRU: Star-shaped Window Transformer Reinforced U-Net for medical image segmentation.

In the last decade, deep neural networks have been widely applied to medical image segmentation, achieving good results in computer-aided diagnosis tasks etc. However, the task of segmenting highly complex, low-contrast images of organs and tissues with high accuracy still faces great challenges. To better address this challenge, this paper proposes a novel model SWTRU (Star-shaped Window Transformer Reinforced U-Net) by combining the U-Net network which plays well in the image segmentation field, and the Transformer which possesses a powerful ability to capture global contexts. Unlike the previous methods that import the Transformer into U-Net, an improved Star-shaped Window Transformer is introduced into the decoder of the SWTRU to enhance the decision-making capability of the whole method. The SWTRU uses a redesigned multi-scale skip-connection model, which retains the inductive bias of the original FCN structure for images while obtaining fine-grained features and coarse-grained semantic information. Our method also presents the FFIM (Filtering Feature Integration Mechanism) to integration and dimensionality reduction of the fused multi-layered features, which reduces the computation. Our SWTRU yields 0.972 DICE on CHLISC for liver and tumor segmentation, 0.897 DICE on LGG for glioma segmentation, and 0.904 DICE on ISIC2018 for skin diseases' segmentation, achieves substantial improvements over the current SoTA across 9 different medical image segment methods. SWTRU can combine feature mapping from different scales, high-level semantics, and global contextual relationships, this architecture is effective in the medical image segmentation. The experimental findings indicate that SWTRU produces superior performance on the medical image segmentation tasks.

Azobenzene Based Photo-Responsive Hydrogel: Synthesis, Self-Assembly, and Antimicrobial Activity.

A new azobenzene-based symmetric amphiphile was synthesized and characterized using 1H NMR spectroscopy. Its self-assembly behavior as well as photo-responsive behavior in its solution and gel states were investigated. Such a compound can self-assemble into fiber mesophases in water solvent. After irradiation of the gels with UV light, the trans isomer of the compound rapidly photoisomerized to the cis isomer, which resulted in a rapid destruction of the gel. High temperature also caused a rapid drop in viscosity. To verify the antimicrobial activity of the hydrogel, live and death assays of human fibroblasts L929 properties were used for in vitro cell viability studies. The compound was converted to the terminal tertiary amine in a quaternary ammonium salt molecule by using hydrochloric acid. This azobenzene quaternary ammonium salt has a relatively better antimicrobial effect biocidal activity that was demonstrated when challenged against Escherichia coli on in vitro conditions.

Insight into the mechanism of allergenicity decreasing in recombinant sarcoplasmic calcium-binding protein from shrimp (Litopenaeus vannamei) with thermal processing via spectroscopy and molecular dynamics simulation techniques.

In the present study, sarcoplasmic calcium-binding protein (SCP) was first expressed in E. coli BL21 (DE3), and then identified based on immunoblotting and SCP amino acid sequencing of shrimp (Litopenaeus vannamei) using mass spectrometry (MS). The recombinant SCP (rSCP) was treated with different temperature conditions to investigate its immunological properties, in vitro digestibility and structural changes with enzyme-linked immunosorbent assay (ELISA), immunoblotting, spectrophotometry and molecular dynamics simulation techniques. The immunoglobulin (Ig) E-binding activity of the rSCP could remain stable until 80 °C, whereas the higher thermal processing temperatures resulted in a significant decrease in IgG/IgE-binding capacity coupled with alterations in the secondary and tertiary structures. Notably, the maximum reduction of IgG/IgE reactivity and in vitro digestibility were observed in the autoclaved rSCP. The decrease in the potential allergenicity of rSCP not only correlated well with the decreasing of α-helix, epitopes masking and exposure of more protease cleavage sites, but also with the destruction of Ca2+ binding sites due to the unfolding of the rSCP with heating treatments, which was supported by the thermal-induced changes of the secondary and tertiary structures. These findings indicate that autoclaved treatment may be an effective and promising approach for producing hypoallergenic seafood.

Discovery of a chalcone derivative as potent necroptosis inhibitor for the treatment of acute kidney injury.

Necroptosis, a form of inflammation-related programmed cell death, is a major mechanism of proximal tubular cell injury in acute kidney injury (AKI). Blockade of necroptosis signalling represents a promising strategy for clinical therapy of AKI. Previously, we identified a small molecular receptor-interacting protein kinases (RIPK)1 inhibitor Cpd-71 with nephroprotective activities. To discover more nephroprotective agents, in this study, 20 chalcone derivatives were synthesized and evaluated for their anti-necroptosis and nephroprotective activities. Among the chalcone derivatives, Cpd-2 exhibited the most potent anti-necroptosis activity (IC50  = 1.08 µM) and protective activity (EC50 = 1.49 µM) through directly binding to RIPK1 and blocking RIPK1-RIPK3-mixed-lineage kinase domain-like protein (MLKL) signalling pathway. Furthermore, Cpd-2 effectively attenuated cisplatin or hypoxia/reoxygenation (H/R)-induced injury and necroptotic inflammation in renal cell models. Moreover, in cisplatin- or ischemia/reperfusion (I/R) induced AKI mouse model, detection of creatinine and urea nitrogen in blood showed that Cpd-2 improved kidney function. Periodic acid-Schiff (PAS) staining and immunofluorescence analysis indicated that Cpd-2 also reduced pathological damage and inhibited inflammatory development in kidney tissues. In summary, although some chalcone derivatives have been reported to prevent kidney injury previously, our present study not only discovered a promising leading compound Cpd-2, but also provided a novel and successful practice for the development of necroptosis inhibitors from natural products derivatives as AKI therapeutic agents.

Thermal induced the structural alterations, increased IgG/IgE binding capacity and reduced immunodetection recovery of tropomyosin from shrimp (Litopenaeus vannamei).

Shrimps were first subjected to various thermal processing, then tropomyosin (TM) was purified and their structure, IgG/IgE-binding ability and detectability were evaluated for elucidating the mechanisms responsible for thermal-induced TM immunodetection recovery alterations. According to CD and FT-IR analysis, heat-treated shrimp TM had significantly reduced α-helix and ß-sheet contents with elevated random coil contents, contributing to an increase of 24.42%-62.22% in IgG/IgE reactivity as compared with raw shrimp TM. The exposure of hydrophobic residues and glycosylation occurred in various heated shrimps TM were confirmed by UV, intrinsic/extrinsic fluorescence spectrum and free amino group analysis, which caused some epitopes masking or modification, thereby inducing considerable TM recovery reduction (48.48%-90.44%). These results demonstrated that thermal-treated TMs with higher structural flexibility facilitated IgG/IgE recognition, however the lower number of epitopes within the thermal-treated TMs might cause considerable underestimation of recovery. The number of antigen binding sites might play a critical role in sandwich immunodetection.

DeepSADPr: A hybrid-learning architecture for serine ADP-ribosylation site prediction.

Protein adenosine diphosphate-ribosylation (ADPr) is caused by the covalent binding of one or more ADP-ribose moieties to a target protein and regulates the biological functions of the target protein. To fully understand the regulatory mechanism of ADP-ribosylation, the essential step is the identification of the ADPr sites from the proteome. As the experimental approaches are costly and time-consuming, it is necessary to develop a computational tool to predict ADPr sites. Recently, serine has been found to be the major residue type for ADP-ribosylation but no predictor is available. In this study, we collected thousands of experimentally validated human ADPr sites on serine residue and constructed several different machine-learning classifiers. We found that the hybrid model, dubbed DeepSADPr, which integrated the one-dimensional convolutional neural network (CNN) with the One-Hot encoding approach and the word-embedding approach, compared favourably to other models in terms of both ten-fold cross-validation and independent test. Its AUC values reached 0.935 for ten-fold cross-validation. Its values of sensitivity, accuracy and Matthews's correlation coefficient reached 0.933, 0.867 and 0.740, respectively, with the fixed specificity value of 0.80. Overall, DeepSADPr is the first classifier for predicting Serine ADPr sites, which is available at http://www.bioinfogo.org/DeepSADPr.

Doping-Induced Quantum Spin Hall Insulator to Superconductor Transition.

A quantum spin Hall insulating state that arises from spontaneous symmetry breaking has remarkable properties: skyrmion textures of the SO(3) order parameter carry charge 2e. Doping this state of matter opens a new route to superconductivity via the condensation of skyrmions. We define a model amenable to large-scale negative sign free quantum Monte Carlo simulations that allows us to study this transition. Our results support a direct and continuous doping-induced transition between the quantum spin Hall insulator and an s-wave superconductor. We can resolve dopings away from half-filling down to δ=0.0017. Such routes to superconductivity have been put forward in the realm of twisted bilayer graphene.

Prediction of arrhythmia after intervention in children with atrial septal defect based on random forest.

BACKGROUND: Using random forest to predict arrhythmia after intervention in children with atrial septal defect. METHODS: We constructed a prediction model of complications after interventional closure for children with atrial septal defect. The model was based on random forest, and it solved the need for postoperative arrhythmia risk prediction and assisted clinicians and patients' families to make preoperative decisions. RESULTS: Available risk prediction models provided patients with specific risk factor assessments, we used Synthetic Minority Oversampling Technique algorithm and random forest machine learning to propose a prediction model, and got a prediction accuracy of 94.65 % and an Area Under Curve value of 0.8956. CONCLUSIONS: Our study was based on the model constructed by random forest, which can effectively predict the complications of arrhythmia after interventional closure in children with atrial septal defect.

LncRNA RP11-248E9.5 and RP11-456D7.1 are Valuable for the Diagnosis of Childhood Pneumonia.

BACKGROUND: Pneumonia is a common infection of the lung parenchyma in children, and early and accurate diagnosis of childhood pneumonia (CP) is important for implementing appropriate preventive and treatment strategies. This study aimed to evaluate the diagnostic value of the combination of long non-coding RNA (lncRNA) RP11-248E9.5, RP11-456D7.1, c-reactive protein (CRP), neutrophil-to-lymphocyte ratio (NLR), and platelet-to-lymphocyte ratio (PLR) in CP. PATIENTS AND METHODS: A total of 50 healthy children (HC) and 100 CP patients were enrolled. The serum expression of RP11-248e9.5 and RP11-456d7.1 was detected by qRT-PCR. The white blood cell (WBC), hemoglobin (HB), platelet (PLT), neutrophil, and lymphocyte were analyzed by automated hematology analyzer. The serum levels of CRP and procalcitonin (PCT) were analyzed by automatic biochemical analyzer. The receiver operating characteristic (ROC) curves were applied to evaluate the diagnostic value in CP. RESULTS: The NLR and PLR, expression of RP11-248E9.5 and RP11-456D7.1, and serum levels of CRP and PCT were significantly higher in the CP group than those in the HC group. Both RP11-248E9.5 (AUC, 0.86; sensitivity, 84%; specificity, 78%) and RP11-456D7.1 (AUC, 0.89; sensitivity, 79%; specificity, 92%) exhibited certain diagnostic value in CP. The diagnostic values of PCT, CRP, NLR and PLR in CP were limited by low sensitivity (≤ 71%). The combination of multiple indicators improved the diagnostic value. The combination of RP11-248E9.5, RP11-456D7.1, CRP, NLR, and PLR had the best diagnostic value in CP (AUC, 0.992; Sensitivity, 0.97; Specificity, 0.99). CONCLUSION: The combination of RP11-248E9.5, RP11-456D7.1, CRP, NLR, and PLR was a potential diagnostic strategy for CP.

Survival analysis for long noncoding RNAs identifies TP53TG1 as an antioncogenic target for the breast cancer.

Breast carcinoma is one of the most commonly diagnosed tumors and also one of the deadliest cancers in the female. Long noncoding RNAs (lncRNAs) are emerging as novel targets and biomarkers for breast cancer diagnosis and treatment. In this study, we aimed to study the lncRNAs associated with the outcomes in patients using the breast invasive carcinoma datasets from The Cancer Genome Atlas. The Cox proportional hazards regression model was fitted to each lncRNA. Hierarchy clustering was carried out using these survival-related lncRNAs and the log-rank test was carried out for the clustered groups. DNA methylation status was utilized to identify the lncRNAs regulated by epigenetics. Finally, the coexpressed messenger RNA with the potential lncRNAs were utilized to study the possible functions and mechanisms of lncRNAs. In total, 182 lncRNAs had an impact on the survival time of the patients with a cutoff <0.01. The patients were clustered into three groups using these survival-related genes, which performed significantly different prognosis. Two lncRNAs, which were significantly correlated with the outcomes of breast cancer and were regulated by methylation status, were obtained. These two lncRNAs were TP53TG1 and RP5-1061H20.4. We proposed that TP53TG1 was activated by the wild-type TP53 and performed an impact on the PI3Ks family by binding YBX2 in breast cancer.

Deep-learning-based information mining from ocean remote-sensing imagery.

With the continuous development of space and sensor technologies during the last 40 years, ocean remote sensing has entered into the big-data era with typical five-V (volume, variety, value, velocity and veracity) characteristics. Ocean remote-sensing data archives reach several tens of petabytes and massive satellite data are acquired worldwide daily. To precisely, efficiently and intelligently mine the useful information submerged in such ocean remote-sensing data sets is a big challenge. Deep learning-a powerful technology recently emerging in the machine-learning field-has demonstrated its more significant superiority over traditional physical- or statistical-based algorithms for image-information extraction in many industrial-field applications and starts to draw interest in ocean remote-sensing applications. In this review paper, we first systematically reviewed two deep-learning frameworks that carry out ocean remote-sensing-image classifications and then presented eight typical applications in ocean internal-wave/eddy/oil-spill/coastal-inundation/sea-ice/green-algae/ship/coral-reef mapping from different types of ocean remote-sensing imagery to show how effective these deep-learning frameworks are. Researchers can also readily modify these existing frameworks for information mining of other kinds of remote-sensing imagery.

Superconductivity from the condensation of topological defects in a quantum spin-Hall insulator.

The discovery of quantum spin-Hall (QSH) insulators has brought topology to the forefront of condensed matter physics. While a QSH state from spin-orbit coupling can be fully understood in terms of band theory, fascinating many-body effects are expected if it instead results from spontaneous symmetry breaking. Here, we introduce a model of interacting Dirac fermions where a QSH state is dynamically generated. Our tuning parameter further allows us to destabilize the QSH state in favour of a superconducting state through proliferation of charge-2e topological defects. This route to superconductivity put forward by Grover and Senthil is an instance of a deconfined quantum critical point (DQCP). Our model offers the possibility to study DQCPs without a second length scale associated with the reduced symmetry between field theory and lattice realization and, by construction, is amenable to large-scale fermion quantum Monte Carlo simulations.

Carbon quantum dots as a fluorophore for "inner filter effect" detection of metronidazole in pharmaceutical preparations.

With houttuynia cordata as carbon source, photoluminescent carbon quantum dots (CDs) were obtained via a one-step hydrothermal procedure. The absorption band of metronidazole (MNZ, maximum absorption wavelength at 319 nm) can well overlap with the excitation bands of CDs (maximum excitation wavelength at 320 nm). A fluorescent approach has been developed for detection of MNZ based on the inner filter effect (IFE), in which as-prepared CDs act as an IFE fluorophore and the MNZ as an IFE absorber. We have investigated the mechanism of quenching the fluorescence of CDs and found that the IFE leads to an exponential decay in fluorescence intensity of CDs with increasing concentration of MNZ, but showed a good linear relationship (R 2 = 0.9930) between ln(F 0/F) with the concentration of MNZ in the range of 3.3 × 10-6 to 2.4 × 10-4 mol L-1. Due to the absence of surface modification of the CDs or establishing any covalent linking between the absorber (MNZ) and the fluorophore (CDs), the developed method is simple, rapid, low-cost and less time-consuming. Meanwhile, it possesses a higher sensitivity, wider linear range, and satisfactory selectivity and has potential application for detection of MNZ in pharmaceutical preparations.

Equivalent-neighbor Potts models in two dimensions.

We investigate the two-dimensional q=3 and 4 Potts models with a variable interaction range by means of Monte Carlo simulations. We locate the phase transitions for several interaction ranges as expressed by the number z of equivalent neighbors. For not-too-large z, the transitions fit well in the universality classes of the short-range Potts models. However, at longer ranges, the transitions become discontinuous. For q=3 we locate a tricritical point separating the continuous and discontinuous transitions near z=80, and a critical fixed point between z=8 and 12. For q=4 the transition becomes discontinuous for z>16. The scaling behavior of the q=4 model with z=16 approximates that of the q=4 merged critical-tricritical fixed point predicted by the renormalization scenario.