Search

Aspect sentiment triple extraction is one of the subtasks of aspect-level sentiment analysis, which aims to extract aspect terms, corresponding opinion terms and sentiment polarity in sentence. Previous studies focus on designing a new paradigm to complete the triplet extraction task in an end-to-end manner. They ignore the role of external knowledge in the model, thus semantic information, part-of-speech information and local context information are not fully explored and utilized. Aiming at the above problems, multi-feature interaction for aspect sentiment triplet extraction (MFI-ASTE) is proposed. Firstly, the bidirectional encoder representation from transformers (BERT) model is used to learn the context semantic feature information, meanwhile, the self-attention mechanism is used to strengthen the semantic feature. Secondly, the semantic feature interacts with the extracted part-of-speech feature and both learn from each other to strengthen the combination ability of the part-of-speech and semantic information. Thirdly, many convolutional neural networks are used to extract multiple local context features of each word, and then multi-point gate mechanism is used to filter these features. Fourthly, three features of external knowledge are fused by two linear layers. Finally, biaffine attention is used for predicting grid tagging and specific decoding schemes are used for decoding triplets. Experimental results show that the proposed model improves the F1 score by 6.83%, 5.60%, 0.54% and 1.22% respectively on four datasets compared with existing mainstream models.

Objective To investigate the expression and clinical significance of H3K27me3 in vascular hemangioma and malformation. Methods Paraffin-embedded tissue blocks from 244 patients with hemangioma and vascular malformation admitted in our hospital from January 2010 and June 2021 were collected. The expression levels of H3K27me3 and smooth muscle actin (SMA) in vascular endothelial and perivascular cells were detected by immunohistochemical SP assay, and its expression differences between hemangioma and vascular malformation was analyzed. Bioinformatics analysis was further applied to explore its potential role in anigodysplasia in combination of the results of SMA immunostaining. Receiver operating characteristic (ROC) curve was drawn to evaluate its diagnostic value in differentiation of hemangioma and vascular malformation. Results In vascular hemangioma, H3K27me3 expression was mainly restricted to vascular endothelial cells, with an expression rate of 50%~90%. However, in vascular malformations, it was expressed in vascular endothelial cells, with an expression rate of 30%~50%, and lost in perivascular cells. There were top 10 endothelium restricted genes derived from GSE21212 dataset based on concerning literatures. According to ChIP-seq data from ENCODE website, H3K27me3 was enriched in the SOX2 promoter of different endothelium lineages, and the molecular interaction of H3K27me3 and SOX2 may be associated with vessel development and differentiation. Conclusion The expression of H3K27me3 in vascular endothelial cells can be used as a candidate marker to distinguish hemangioma and vascular malformation. However, its expression in infantile hemangioma is immunologically heterogeneous.

Objective To detect the expression of 4 progesterone receptor (PR) antibodies (EP2, SP2, 1E2, and 16) in non-specific invasive breast cancer (NSIBCs) and identify the optimal antibody for PR analysis and treatment planning for NSIBCs. Methods Immunohistochemistry was used to detect the expression of the 4 PR antibodies in a tissue microarray consisting of 195 tissue specimens of NSIBCs, and their constituent ratios with the expression of estrogen receptor (ER), the percentages of stained cells, staining intensity, Allred score (A-score) and consistency were compared. Results The constituent ratios of the 4 PR antibodies with ER (ER±/PR±) did not differ significantly in the 195 cases of NSIBCs (P>0.05). Regardless of the status of ER expression, the cases showed high rates of positivity or negativity for all the 4 PR antibodies at the same time (81.54%~90.77%) with good consistencies among the 4 antibodies (κ=0.634~0.804). Further analysis of the average number of stained cells, percentage of positive cells, staining intensity, and A-score of the 4 PR antibodies showed that all these indicators were significantly higher for 1E2 than for EP2, SP2, and 16 (P < 0.05), and were not significantly different among the latter 3 antibodies (P>0.05). Conclusion 1E2 has a higher positivity rate and expression intensity in NSIBCs than EP2, SP2 and RP16, and thus can serve as an ideal PR antibody for clinical pathology tests.

Background: Inflammation is the common pathogenesis of coronary atherosclerosis disease (CAD) and rheumatoid arthritis (RA). Although it is established that RA increases the risk of CAD, the underlining mechanism remained indefinite. This study seeks to explore the molecular mechanisms of RA linked CAD and identify potential target gene for early prediction of CAD in RA patients. Materials and methods: The study utilized five raw datasets: GSE55235, GSE55457, GSE12021 for RA patients, and GSE42148 and GSE20680 for CAD patients. Gene Set Enrichment Analysis (GSEA) was used to investigate common signaling pathways associated with RA and CAD. Then, weighted gene co-expression network analysis (WGCNA) was performed on RA and CAD training datasets to identify gene modules related to single-sample GSEA (ssGSEA) scores. Overlapping module genes and differentially expressed genes (DEGs) were considered as co-susceptible genes for both diseases. Three hub genes were screened using a protein-protein interaction (PPI) network analysis via Cytoscape plug-ins. The signaling pathways, immune infiltration, and transcription factors associated with these hub genes were analyzed to explore the underlying mechanism connecting both diseases. Immunohistochemistry and qRT-PCR were conducted to validate the expression of the key candidate gene, PPARG, in macrophages of synovial tissue and arterial walls from RA and CAD patients. Results: The study found that Fc-gamma receptor-mediated endocytosis is a common signaling pathway for both RA and CAD. A total of 25 genes were screened by WGCNA and DEGs, which are involved in inflammation-related ligand-receptor interactions, cytoskeleton, and endocytosis signaling pathways. The principal component analysis(PCA) and support vector machine (SVM) and receiver-operator characteristic (ROC) analysis demonstrate that 25 DEGs can effectively distinguish RA and CAD groups from normal groups. Three hub genes TUBB2A, FKBP5, and PPARG were further identified by the Cytoscape software. Both FKBP5 and PPARG were downregulated in synovial tissue of RA and upregulated in the peripheral blood of CAD patients and differential mRNAexpreesion between normal and disease groups in both diseases were validated by qRT-PCR.Association of PPARG with monocyte was demonstrated across both training and validation datasets in CAD. PPARG expression is observed in control synovial epithelial cells and foamy macrophages of arterial walls, but was decreased in synovial epithelium of RA patients. Its expression in foamy macrophages of atherosclerotic vascular walls exhibits a positive correlation (r = 0.6276, p = 0.0002) with CD68. Conclusion: Our findings suggest that PPARG may serve as a potentially predictive marker for CAD in RA patients, which provides new insights into the molecular mechanism underling RA linked CAD. [ABSTRACT FROM AUTHOR]

Pancreatic intraepithelial neoplasia (PanIN) is the most common precursor lesion that has the potential to progress to invasive pancreatic cancer, and early and rapid detection may offer patients a chance for treatment before the development of invasive carcinoma. Therefore, the identification of PanIN holds significant clinical importance. In this study, we first used multiphoton microscopy (MPM) combining two‐photon excitation fluorescence and second‐harmonic generation imaging to label‐free detect PanIN and attempted to differentiate between normal pancreatic ducts and different grades of PanIN. Then, we also developed an automatic image processing strategy to extract eight morphological features of collagen fibers from MPM images to quantify the changes in collagen fibers surrounding the ducts. Experimental results demonstrate that the combination of MPM and quantitative information can accurately identify normal pancreatic ducts and different grades of PanIN. This study may contribute to the rapid diagnosis of pancreatic diseases and may lay the foundation for further clinical application of MPM. [ABSTRACT FROM AUTHOR]