课程文献中心 更多>>

关键词： Camellia sinensis L   Spatial heterogeneity   Flavor formation   Consumer preference   Raw material harvesting and grading   Quality evaluation   AAs   Amino Acids   GA   Gallic Acid   EGCG   Epigallocatechin Gallate   EGC   Epigallocatechin   GCG   Gallocatechin Gallate   GC   Gas Chromatography   LC   liquid chromatography   SICRIT   Soft Ionization by Chemical Reaction in Transfer   SEM   Scanning Electron Microscopy   SVM   Support Vector Machine   SERS   Surface-enhanced Raman Spectroscopy   CAM   Global Attention Mechanism   SVR   Support Vector Regression   GABA   γ-Aminobutyric Acid   PRPs   Proline-Rich Proteins   ECG   Epicatechin Gallate   EC   Epicatechin   CG   Catechin Gallate   MS   Mass Spectrometry   DART   Direct Analysis in Real Time   PLSR   Partial Least Squares Regression   TDS   Temporal Dominance of Sensations   RF   Random Forest   CNN   Convolutional Neural Network   UHPLC   Ultra-High Performance Liquid Chromatography   NIR   Near-Infrared  

摘要：

• •
  A systematic framework of spatial flavor metabolism in tea shoots is established.
• •
  Metabolic specialization of tea shoot tissues is comprehensively summarized.
• •
  Spatial metabolic interactions governing tea flavor are critically discussed.
• •
  The review provides guidance for precision tea production and quality evaluation.

关键词： Multiple privileged information   Dual-level transfer   Image-only inference   Self-generated visual features   CLIP-based alignment  

摘要： Computer-aided diagnosis (CAD) generally cannot sufficiently capture knowledge from images. In clinical practice, textual reports accompany medical images and contain rich diagnostic descriptions, which have potential to provide valuable supplementary information to enhance the image-based CAD. However, such textual data are unavailable during model inference in real-world clinical workflow. To address this challenge, a novel Support Vector Machine Plus (SVM+) based bidirectional generalized distillation (BGD) for multiple privileged information (MPI-SVM+-BGD) framework is proposed under the learning using privileged information (LUPI) paradigm. In particular, the framework leverages textual semantic embeddings from clinical reports and selfgenerated visual structural features from images as dual privileged information (PI). The SVM+-BGD framework integrates theSVM+ classifier with the BGD framework to establishes a dual-level transfer mechanism that operates at both the classifier-level and the logit-level. Moreover, a variant of SVM+-BGD for Multiple PIs, named MPI-SVM+-BGD, is further developed for jointly optimizing both the textual and self-generated visual PIs. This framework enables multimodal knowledge transfer learning during model training while maintaining unimodal image-only inference. Extensive experiments on breast ultrasound and gastric histopathology datasets demonstrate that the proposed MPI-SVM+-BGD outperforms all the compared methods while maintaining an imageonly inference pipeline fully consistent with clinical workflows.

关键词： CLIP-based alignment   Dual-level transfer   Image-only inference   Multiple privileged information   Self-generated visual features  

摘要： Computer-aided diagnosis (CAD) generally cannot sufficiently capture knowledge from images. In clinical practice, textual reports accompany medical images and contain rich diagnostic descriptions, which have potential to provide valuable supplementary information to enhance the image-based CAD. However, such textual data are unavailable during model inference in real-world clinical workflow. To address this challenge, a novel Support Vector Machine Plus (SVM+) based bidirectional generalized distillation (BGD) for multiple privileged information (MPI-SVM+-BGD) framework is proposed under the learning using privileged information (LUPI) paradigm. In particular, the framework leverages textual semantic embeddings from clinical reports and self-generated visual structural features from images as dual privileged information (PI). The SVM+-BGD framework integrates theSVM+ classifier with the BGD framework to establishes a dual-level transfer mechanism that operates at both the classifier-level and the logit-level. Moreover, a variant of SVM+-BGD for Multiple PIs, named MPI-SVM+-BGD, is further developed for jointly optimizing both the textual and self-generated visual PIs. This framework enables multimodal knowledge transfer learning during model training while maintaining unimodal image-only inference. Extensive experiments on breast ultrasound and gastric histopathology datasets demonstrate that the proposed MPI-SVM+-BGD outperforms all the compared methods while maintaining an image-only inference pipeline fully consistent with clinical workflows.

关键词： Deep learning   Feature extraction   Fourier transform infrared spectroscopy   Gradient-weighted class activation mapping   Microplastic identification  

摘要： To address the low accuracy of spectral identification of mixed microplastic samples and the “black box” nature of deep learning models, this paper proposes an optimization method based on Gradient Weighted Class Activation Mapping (Grad-CAM). Model interpretability analysis using Grad-CAM reveals an imbalance in the feature extraction process—insufficient attention to the features of minor components in mixed samples. Based on these findings, multi-head attention and multi-region attention mechanisms are introduced to enhance the model's ability to extract global and local features. Considering the complex and diverse ATR-FTIR spectral features of microplastics and the gradient vanishing problem in deep learning models, Resblock residual connections are introduced on top of CNN, and a ResM2A-CNN model is designed. Experimental results show that the overall accuracy of the optimized model on the test set is improved from 95.67% to 99.92%. Notably, the ResM2A-CNN model exhibits strong classification ability for PA and its mixtures, outperforming other traditional models in various performance metrics.This study highlights the effectiveness of using interpretability analysis to guide model optimization, providing a breakthrough method for the rapid and accurate detection of microplastics. Furthermore, this research offers practical insights into modeling complex spectral data and optimizing deep learning models.

关键词： Protein disulfide isomerase   PDIA1   PDIA3   Small molecule inhibitors   Tetrahydro-β-carboline derivatives   PDIs   Protein disulfide isomerases   TRX-like   thioredoxin-like   ER   endoplasmic reticulum   CNX   calnexin   CRT   calreticulin   HA   hemagglutinin   THBC   tetrahydro-β-carboline   DSF   differential scanning fluorimetry   IFD   induced-fit docking   TEA   triethylamine   SASA   Solvent Accessible Surface Area   MD   molecular dynamics   ESP   Electrostatic Potential   QM   quantum mechanics   FWHM   full width at half maximum   CAM   carbamidomethylation  

摘要：

• •
  New tetrahydro-β-carbolines synthesized as PDIA1 and PDIA3 inhibitors.
• •
  Compound 2e showed promising biological activities in vitro and in cellulo assays.
• •
  Proteomic analysis showed a distinct Cys modification profile in PDIA1 and PDIA3.
• •
  Molecular modeling mapped binding modes and thiol reactivity to key cysteines.

关键词： Hand–eye calibration   Simultaneous calibration   Gauss–Newton method   Robotic measurement  

摘要：

• •
  A unified hand–eye calibration model is established via point cloud registration and motion chain modeling.
• •
  A two-stage calibration strategy combining closed-form initialization and iterative refinement is proposed.
• •
  The proposed method can achieve online calibration through general objects.
• •
  Measurement and peg-in-hole experiments demonstrate improved accuracy and stability.

关键词： Symbolic regression   CAD-free   Lifting surface   Mass modeling   Volume computation   Deep neural networks  

摘要： Accurate structural mass estimation of lifting surfaces is required throughout aircraft design, yet existing methods cannot simultaneously provide closed-form geometric descriptors, physically consistent similarity scaling, and exact differentiability without repeated three-dimensional geometry instantiation. This work presents a symbolic-regression (SR) framework for segment-level mass estimation of solid, skin-laminated, shell, and wing-box lifting surfaces, using external surface area and enclosed volume as governing mass descriptors and quasi-analytical rib and spar volume expressions for internal contributions.

关键词： Digital PCR   Mitochondrial DNA   Chicken chorioallantoic membrane (CAM) assay   Metastatic burden   Patient-derived xenograft (PDX)  

摘要：

• •
  The chorioallantoic membrane (CAM) assay facilitates tumor metastasis research.
• •
  Quantifying this process via Alu-qPCR suffers from false positive signaling.
• •
  Human mitochondrial MT-ND4 offers a highly specific alternative target.
• •
  MT-ND4 digital PCR allows sensitive human DNA detection in the CAM assay.

关键词： Internal dosimetry   Whole body counter   MCNP6   AutoCAD  

摘要：

• •
  The use of CAD software to generate model geometry for MCNP.
• •
  Validation of the simulated model, with a variation of 9.4% compared to the experimental data.
• •
  The validated model enabled the first case study, with preliminary conclusions consistent with the expected results.

关键词： AAA   abdominal aortic aneurysm   ACKRs   atypical chemokine receptors   ACS   acute coronary syndrome   AKT   protein kinase B   AMI   acute myocardial infarction   Ang II   angiotensin II   AP-1   activator protein 1   Apo E   apolipoprotein E-knockout   AQP4   aquaporin 4   ASAH   aneurysmal subarachnoid hemorrhage   AS-IV   astragaloside IV   BBB   blood-brain barrier   BIN   bindarit   BM   bone marrow   CA   cerebral aneurysm   CAD   coronary artery disease   CAO   coronary artery occlusion   CaSR   calcium-sensing receptor   CCL5   CC motif ligand 5   CCR1   CC motif chemokine receptor 1   CDC1   CD11b DCs   CIRI   cerebral ischemia-reperfusion injury   CREB   cAMP response element binding protein   DCM   dilated cardiomyopathy   DCs   dendritic cells   DOCA   deoxycorticosterone acetate   DPP-IV   dipeptidyl peptidase IV   ECM   extracellular matrix   EPCs   endothelial progenitor cells   ERK1   extracellular signal-regulated kinase 1   ESCC   esophageal squamous cell carcinoma   FOXO1   forehead box O1   FPRL-1   formyl peptide receptor-like-1   GAGs   glycosaminoglycans   GCA   giant cell arteritis   GMCSF   granulocyte-macrophage colony-stimulating factor   GPCRs   Gαi-protein-coupled seven-transmembrane-spanning receptors   HCM   hypertrophic cardiomyopathy   HDL   high density lipoprotein   HSC   hepatic stellate cells   HUC-MSCs   human umbilical cord-derived MSCs   ICH   intracerebral hemorrhage   IFN-γ   interferon-gamma   IL-4   interleukin-4   IVH   intraventricular hemorrhage   INOS   inducible nitric oxide synthase   JAK2   janus kinase 2   JNK   c-Jun N-terminus kinase   LARC   liver- and activation-regulated chemokine   LncRNAs   long non-coding RNAs   LPS   lipopolysaccharide   LSS   low shear stress   MAPK   mitogen-activated protein kinase   MCAO   middle cerebral artery occlusion   MCP   monocytes chemoattractant protein   MCPIP-1   monocyte chemotactic protein-1-induced protein-1   MDMs   monocyte-derived macrophages   Met-CCL5   methionylated CCL5   MI   myocardial infarction   MIP   macrophages inflammatory protein-1   MIRI   myocardial ischemia-reperfusion injury   MiRNAs   microRNAs   MMPs   matrix metalloproteinases   MSCs   mesenchymal stem cell   MTOR   mammalian target of rapamycin   MVC   maraviroc   NcRNAs   non-coding RNAs   NF-κB   nuclear  

摘要： The CC chemokine family is a group of low-molecular-weight secreted proteins that can induce various cells to migrate to specific tissues. CC chemokines are also coordinators of many cellular activities, including hematopoiesis, wound healing, tumor growth and metastasis, fibrosis, etc. Recently, CC chemokines play a pleiotropic role in regulating the occurrence and progression of cardiovascular and cerebrovascular diseases, mainly by exacerbating the diseases, and their effects vary at different stages of disease progression. Moreover, the expression of many CC chemokines is highly specific in cardiovascular and cerebrovascular diseases. These CC chemokines can serve as potential biomarkers for predicting future cardiovascular events. Intervention strategies targeting CC chemokines to alleviate diseases, such as blocking the interaction between CC chemokines and their receptors, interfering with the binding of chemokines with glycosaminoglycans, have shown great potential for application. In this review, we present the structure, physiological functions, and regulatory mechanisms of CC family, particularly elaborate on the association between CC chemokines and cardiovascular and cerebrovascular diseases, and reviewing the therapeutic applications of CC chemokines. This review demonstrates the great potential of CC chemokines in the treatment of cardiovascular and cerebrovascular diseases and proposes directions for future research.