Body composition trajectories during childhood predict skeletal maturation at puberty: A longitudinal study.

Nutritional status significantly impacts linear bone growth. We aimed to determine the relationship between the trajectories of four body composition indicators and pubertal advanced bone age. Trajectories of body mass index z-score (BMI z-score), visceral fat area z-score (VFA z-score), fat mass index z-score (FMI z-score), and fat-free mass index z-score (FFMI z-score) were identified based on three body composition measurements conducted from October 2018 to April 2023 within a pediatric cohort (the PROC study). We assessed pubertal bone age using the Tanner-Whitehouse 3-Chinese Radius-Ulna-Short (TW3-C RUS) method among 1402 primary school children. Children with a trajectory of higher BMI z-score, VFA z-score, FMI z-score, and FFMI z-score since childhood were more likely to have advanced bone age. The risk of advanced bone age was higher in children who were consistently in the high VFA z-score group (odds ratio [OR] = 6.73) or consistently in the high BMI z-score group (OR = 5.57), as compared to those in the low VFA z-score and low BMI z-score groups. Regular monitoring and maintenance of normal VFA during childhood may reduce the risk of advanced bone age at puberty. Furthermore, BMI monitoring is optional, especially in cases where specialized body composition equipment is not available.

Prediction of Mechanical Properties of Lattice Structures: An Application of Artificial Neural Networks Algorithms.

The yield strength and Young's modulus of lattice structures are essential mechanical parameters that influence the utilization of materials in the aerospace and medical fields. Currently, accurately determining the Young's modulus and yield strength of lattice structures often requires conduction of a large number of experiments for prediction and validation purposes. To save time and effort to accurately predict the material yield strength and Young's modulus, based on the existing experimental data, finite element analysis is employed to expand the dataset. An artificial neural network algorithm is then used to establish a relationship model between the topology of the lattice structure and Young's modulus (the yield strength), which is analyzed and verified. The Gibson-Ashby model analysis indicates that different lattice structures can be classified into two main deformation forms. To obtain an artificial neural network model that can accurately predict different lattice structures and be deployed in the prediction of BCC-FCC lattice structures, the artificial network model is further optimized and validated. Concurrently, the topology of disparate lattice structures gives rise to a certain discrete form of their dominant deformation, which consequently affects the neural network prediction. In conclusion, the prediction of Young's modulus and yield strength of lattice structures using artificial neural networks is a feasible approach that can contribute to the development of lattice structures in the aerospace and medical fields.

National age-specific mortality trends for cervical and breast cancers in urban-rural areas of China from 2009 to 2021: a population-based analysis.

BACKGROUND: Cervical and breast cancers are among the top 4 leading causes of cancer-related mortality in women. This study aimed to examine age-specific temporal trends in mortality for cervical and breast cancers in urban and rural areas of China from 2009 to 2021. METHODS: Age-specific mortality data for cervical and breast cancers among Chinese women aged 20-84 years were obtained from China's National Disease Surveillance Points system spanning the years 2009 to 2021. Negative binomial regression models were utilized to assess urban-rural differences in mortality rate ratios, while Joinpoint models with estimated average annual percent changes (AAPC) and slopes were employed to compare temporal trends and the acceleration of mortality rates within different age groups. RESULTS: From 2009 to 2021, there was a relative increase in age-specific mortality associated with the two cancers observed in rural areas compared with urban areas. A rising trend in the screening age of 35-64 [AAPC: 4.0%, 95% confidence interval (CI) 0.5-7.6%, P = 0.026] for cervical cancer was noted in rural areas, while a stable trend (AAPC: - 0.7%, 95% CI - 5.8 to 4.6%, P = 0.78) was observed in urban areas. As for breast cancer, a stable trend (AAPC: 0.3%, 95% CI - 0.3 to 0.9%, P = 0.28) was observed in rural areas compared to a decreasing trend (AAPC: - 2.7%, 95% CI - 4.6 to - 0.7%, P = 0.007) in urban areas. Urban-rural differences in mortality rates increased over time for cervical cancer but decreased for breast cancer. Mortality trends for both cervical and breast cancers showed an increase with age across 4 segments, with the most significant surge in mortality observed among the 35-54 age group across urban and rural areas, periods, and regions in China. CONCLUSIONS: Special attention should be given to women aged 35-54 years due to mortality trends and rural-urban disparities. Focusing on vulnerable age groups and addressing rural-urban differences in the delivery of cancer control programs can enhance resource efficiency and promote health equity.

Metabolically healthy obesity and left ventricular geometric remodelling in Chinese children.

AIM: To investigate the association between metabolically healthy obesity (MHO) and left ventricular geometric remodelling in Chinese children. MATERIALS AND METHODS: This cross-sectional study used data from two population-based samples in China, including 2871 children aged 6-11 years. Weight status was defined based on body mass index according to the World Health Organization growth chart. Metabolic status was defined based on the 2018 consensus-based criteria proposed by Damanhoury et al. Obes Rev 2018;19:1476-1491 (blood pressure, lipids and glucose). Left ventricular geometric remodelling was determined as concentric remodelling, eccentric hypertrophy, and concentric hypertrophy. Multinomial logistic regression analysis was used to determine odds ratios (ORs) and 95% confidence intervals (CIs) for the association between categories of weight and metabolic status and left ventricular geometric remodelling. RESULTS: Compared with children with metabolically healthy normal weight, those with MHO had higher odds of left ventricular geometric remodelling, with adjusted ORs (95% CIs) of 2.01 (1.23-3.28) for concentric remodelling, 6.36 (4.03-10.04) for eccentric hypertrophy, and 17.07 (7.97-36.58) for concentric hypertrophy. Corresponding ORs (95% CIs) were 2.35 (1.47-3.75), 10.85 (7.11-16.55), and 18.56 (8.63-39.94), respectively, for children with metabolically unhealthy obesity. In contrast, metabolically unhealthy normal weight was not associated with higher odds of left ventricular geometric remodelling. Findings were consistent in sensitivity analyses that used different definitions of weight and metabolic status and left ventricular geometric remodelling. CONCLUSIONS: Children with MHO had higher odds of left ventricular geometric remodelling than their metabolically healthy normal weight counterparts. Our findings suggest MHO may not be a benign condition for cardiac health in children.

Longitudinal mass loads of phthalate esters in sewage sludge and their association with public health and social measures for COVID-19 control in Beijing.

This study examines the concentrations and population-normalized mass loads (PNML) of five phthalate esters (PAEs) and four metabolites (mPAEs) in 390 sewage sludge samples collected from two municipal wastewater treatment plants in Beijing between July 2020 and June 2023, amidst the COVID-19 pandemic. Through GC/MS analysis, the compounds were simultaneously quantified, with peak concentrations in 2020. Bis(2-ethylhexyl) phthalate (DEHP) and mono(2-ethyl-5-oxohexyl) phthalate emerged as predominant PAE and mPAE congeners with concentrations of 78.7 µg/g dw and 259 µg/g dw, respectively. DEHP and monobenzyl phthalate had the highest median PNML among PAEs and mPAEs, respectively, at 128 µg/inhabitant/day and 798 µg/inhabitant/day. Linear regression models revealed a positive association between PNML of PAEs and five public health and social measures aimed at mitigating the COVID-19 pandemic. This research contributes to the expanding body of literature by emphasizing the role of wastewater-based epidemiology as a vital tool for monitoring community-level exposure to environmental contaminants.

A Polygenic Risk Analysis for Identifying Ulcerative Colitis Patients with European Ancestry.

The incidence of ulcerative colitis (UC) has increased globally. As a complex disease, the genetic predisposition for UC could be estimated by the polygenic risk score (PRS), which aggregates the effects of a large number of genetic variants in a single quantity and shows promise in identifying individuals at higher lifetime risk of UC. Here, based on a cohort of 2869 UC cases and 2900 controls with genotype array datasets, we used PRSice-2 to calculate PRS, and systematically analyzed factors that could affect the power of PRS, including GWAS summary statistics, population stratification, and impact of variants. After leveraging a stepwise condition analysis, we eventually established the best PRS model, achieving an AUC of 0.713. Meanwhile, samples in the top 20% of the PRS distribution had a risk of UC more than ten times higher than samples in the lowest 20% (OR = 10.435, 95% CI 8.571-12.703). Our analyses demonstrated that including population-enriched, more disease-associated SNPs and using GWAS summary statistics from similar ethnic background can improve the power of PRS. Strictly following the principle of focusing on one population in all aspects of generating PRS can be a cost-effective way to apply genotype-array-derived PRS to practical risk estimation.

Photochemical dearomative skeletal modifications of heteroaromatics.

Dearomatization has emerged as a powerful tool for rapid construction of 3D molecular architectures from simple, abundant, and planar (hetero)arenes. The field has evolved beyond simple dearomatization driven by new synthetic technology development. With the renaissance of photocatalysis and expansion of the activation mode, the last few years have witnessed impressive developments in innovative photochemical dearomatization methodologies, enabling skeletal modifications of dearomatized structures. They offer truly efficient and useful tools for facile construction of highly complex structures, which are viable for natural product synthesis and drug discovery. In this review, we aim to provide a mechanistically insightful overview on these innovations based on the degree of skeletal alteration, categorized into dearomative functionalization and skeletal editing, and to highlight their synthetic utilities.

Reversing Metabolic Dysfunction-Associated Steatotic Liver Disease Promotes Healthy Pediatric Cardiovascular Structures: The PROC Study.

INTRODUCTION: We investigated the impact of metabolic dysfunction-associated steatotic liver disease (MASLD) on cardiovascular structure development in children. METHODS: We followed 1,356 children with the mean age of 6.6 years for 4.5 years in Beijing, China. We assessed the association of MASLD with cardiovascular structure (carotid intima-media thickness and left ventricular mass) outcomes at baseline and follow-up. RESULTS: Over follow-up, 59 children had persistent MASLD, 109 had incident MASLD (progression), and 35 had normalization of liver health. Children with MASLD normalization showed a significantly lower mean development in carotid intima-media thickness (0.161 vs 0.188 mm) and left ventricular mass (4.5 vs 12.4 g) than children with persistent MASLD. DISCUSSION: The control of MASLD was associated with improved cardiovascular structure development.

Advancing Adverse Drug Reaction Prediction with Deep Chemical Language Model for Drug Safety Evaluation.

The accurate prediction of adverse drug reactions (ADRs) is essential for comprehensive drug safety evaluation. Pre-trained deep chemical language models have emerged as powerful tools capable of automatically learning molecular structural features from large-scale datasets, showing promising capabilities for the downstream prediction of molecular properties. However, the performance of pre-trained chemical language models in predicting ADRs, especially idiosyncratic ADRs induced by marketed drugs, remains largely unexplored. In this study, we propose MoLFormer-XL, a pre-trained model for encoding molecular features from canonical SMILES, in conjunction with a CNN-based model to predict drug-induced QT interval prolongation (DIQT), drug-induced teratogenicity (DIT), and drug-induced rhabdomyolysis (DIR). Our results demonstrate that the proposed model outperforms conventional models applied in previous studies for predicting DIQT, DIT, and DIR. Notably, an analysis of the learned linear attention maps highlights amines, alcohol, ethers, and aromatic halogen compounds as strongly associated with the three types of ADRs. These findings hold promise for enhancing drug discovery pipelines and reducing the drug attrition rate due to safety concerns.

Identifying diagnostic indicators for type 2 diabetes mellitus from physical examination using interpretable machine learning approach.

Background: Identification of patients at risk for type 2 diabetes mellitus (T2DM) can not only prevent complications and reduce suffering but also ease the health care burden. While routine physical examination can provide useful information for diagnosis, manual exploration of routine physical examination records is not feasible due to the high prevalence of T2DM. Objectives: We aim to build interpretable machine learning models for T2DM diagnosis and uncover important diagnostic indicators from physical examination, including age- and sex-related indicators. Methods: In this study, we present three weighted diversity density (WDD)-based algorithms for T2DM screening that use physical examination indicators, the algorithms are highly transparent and interpretable, two of which are missing value tolerant algorithms. Patients: Regarding the dataset, we collected 43 physical examination indicator data from 11,071 cases of T2DM patients and 126,622 healthy controls at the Affiliated Hospital of Southwest Medical University. After data processing, we used a data matrix containing 16004 EHRs and 43 clinical indicators for modelling. Results: The indicators were ranked according to their model weights, and the top 25% of indicators were found to be directly or indirectly related to T2DM. We further investigated the clinical characteristics of different age and sex groups, and found that the algorithms can detect relevant indicators specific to these groups. The algorithms performed well in T2DM screening, with the highest area under the receiver operating characteristic curve (AUC) reaching 0.9185. Conclusion: This work utilized the interpretable WDD-based algorithms to construct T2DM diagnostic models based on physical examination indicators. By modeling data grouped by age and sex, we identified several predictive markers related to age and sex, uncovering characteristic differences among various groups of T2DM patients.

Geraniol attenuates oxygen-glucose deprivation/reoxygenation-induced ROS-dependent apoptosis and permeability of human brain microvascular endothelial cells by activating the Nrf-2/HO-1 pathway.

Blood-brain barrier breakdown and ROS overproduction are important events during the progression of ischemic stroke aggravating brain damage. Geraniol, a natural monoterpenoid, possesses anti-apoptotic, cytoprotective, anti-oxidant, and anti-inflammatory activities. Our study aimed to investigate the effect and underlying mechanisms of geraniol in oxygen-glucose deprivation/reoxygenation (OGD/R)-induced human brain microvascular endothelial cells (HBMECs). Apoptosis, caspase-3 activity, and cytotoxicity of HBMECs were evaluated using TUNEL, caspase-3 activity, and CCK-8 assays, respectively. The permeability of HBMECs was examined using FITC-dextran assay. Reactive oxygen species (ROS) production was measured using the fluorescent probe DCFH-DA. The protein levels of zonula occludens-1 (ZO-1), occludin, claudin-5, ß-catenin, nuclear factor erythroid 2-related factor 2 (Nrf2), and heme oxygenase-1 (HO-1) were determined by western blotting. Geraniol showed no cytotoxicity in HBMECs. Geraniol and ROS scavenger N-acetylcysteine (NAC) both attenuated OGD/R-induced apoptosis and increase of caspase-3 activity and the permeability to FITC-dextran in HBMECs. Geraniol relieved OGD/R-induced ROS accumulation and decrease of expression of ZO-1, occludin, claudin-5, and ß-catenin in HBMECs. Furthermore, we found that geraniol activated Nrf2/HO-1 pathway to inhibit ROS in HBMECs. In conclusion, geraniol attenuated OGD/R-induced ROS-dependent apoptosis and permeability in HBMECs through activating the Nrf2/HO-1 pathway.

Erratum: FAM72A promotes glioma progression by regulating mitophagy through the Pink1/Parkin signaling pathway: Erratum.

[This corrects the article DOI: 10.7150/jca.82949.].

Wide-range soft anisotropic thermistor with a direct wireless radio frequency interface.

Temperature sensors are one of the most fundamental sensors and are found in industrial, environmental, and biomedical applications. The traditional approach of reading the resistive response of Positive Temperature Coefficient thermistors at DC hindered their adoption as wide-range temperature sensors. Here, we present a large-area thermistor, based on a flexible and stretchable short carbon fibre incorporated Polydimethylsiloxane composite, enabled by a radio frequency sensing interface. The radio frequency readout overcomes the decades-old sensing range limit of thermistors. The composite exhibits a resistance sensitivity over 1000 °C-1, while maintaining stability against bending (20,000 cycles) and stretching (1000 cycles). Leveraging its large-area processing, the anisotropic composite is used as a substrate for sub-6 GHz radio frequency components, where the thermistor-based microwave resonators achieve a wide temperature sensing range (30 to 205 °C) compared to reported flexible temperature sensors, and high sensitivity (3.2 MHz/°C) compared to radio frequency temperature sensors. Wireless sensing is demonstrated using a microstrip patch antenna based on a thermistor substrate, and a battery-less radio frequency identification tag. This radio frequency-based sensor readout technique could enable functional materials to be directly integrated in wireless sensing applications.

HOPEXGB: A Consensual Model for Predicting miRNA/lncRNA-Disease Associations Using a Heterogeneous Disease-miRNA-lncRNA Information Network.

Predicting disease-related microRNAs (miRNAs) and long noncoding RNAs (lncRNAs) is crucial to find new biomarkers for the prevention, diagnosis, and treatment of complex human diseases. Computational predictions for miRNA/lncRNA-disease associations are of great practical significance, since traditional experimental detection is expensive and time-consuming. In this paper, we proposed a consensual machine-learning technique-based prediction approach to identify disease-related miRNAs and lncRNAs by high-order proximity preserved embedding (HOPE) and eXtreme Gradient Boosting (XGB), named HOPEXGB. By connecting lncRNA, miRNA, and disease nodes based on their correlations and relationships, we first created a heterogeneous disease-miRNA-lncRNA (DML) information network to achieve an effective fusion of information on similarities, correlations, and interactions among miRNAs, lncRNAs, and diseases. In addition, a more rational negative data set was generated based on the similarities of unknown associations with the known ones, so as to effectively reduce the false negative rate in the data set for model construction. By 10-fold cross-validation, HOPE shows better performance than other graph embedding methods. The final consensual HOPEXGB model yields robust performance with a mean prediction accuracy of 0.9569 and also demonstrates high sensitivity and specificity advantages compared to lncRNA/miRNA-specific predictions. Moreover, it is superior to other existing methods and gives promising performance on the external testing data, indicating that integrating the information on lncRNA-miRNA interactions and the similarities of lncRNAs/miRNAs is beneficial for improving the prediction performance of the model. Finally, case studies on lung, stomach, and breast cancers indicate that HOPEXGB could be a powerful tool for preclinical biomarker detection and bioexperiment preliminary screening for the diagnosis and prognosis of cancers. HOPEXGB is publicly available at https://github.com/airpamper/HOPEXGB.

Relationship among physical exercise, social support and sense of coherence in rural left-behind children.

OBJECTIVE: Sense of coherence has significant implications for the mental health of left-behind children in rural China. This study was conducted to explore the impact of physical exercise and social support on sense of coherence in rural left-behind children. METHODS: A survey was conducted among 964 rural left-behind children in Hunan Province of China aged 10-15 (Mage = 12.01 ± 1.97 years; 529 girls and 435 boys). AMOS 21.0 software was applied to build the model and perform cross-lagged analysis. RESULTS: Physical exercise and social support in left-behind children predict each other, and physical exercise can positively predict social support and sense of coherence in left-behind children two months later (P < 0.05); social support can significantly positively predict sense of coherence in left-behind children two months later (P < 0.05). CONCLUSIONS: Both physical exercise and social support have an impact on sense of coherence in left-behind children, and physical exercise can affect sense of coherence in left-behind children through the mediating effect of social support.

Comparison of Glenohumeral Bone Morphology Between Patients With Versus Without Anterior Shoulder Instability.

Background: The stability of the glenohumeral joint is associated with anatomic characteristics including bony structures and soft tissues. Purpose: To compare the differences in specific bony glenohumeral geometries between shoulders with anterior shoulder instability (ASI), unaffected contralateral shoulders, and healthy control shoulders. Study Design: Cross-sectional study; Level of evidence, 3. Methods: Shoulder computed tomography (CT) scans of 36 patients with ASI and 36 matched healthy controls were retrieved and 3-dimensionally reconstructed. We measured the glenoid radius of curvature (GROC) in the anterior-posterior (AP) and superior-inferior directions, humeral head radius of curvature (HROC) in the AP direction, conformity index, glenoid height, glenoid width, glenoid index, stability angle, glenoid version, and glenoid depth. The differences between the groups were statistically calculated. CT scans of the unaffected contralateral shoulders from 21 of the ASI patients were also collected to identify the consistency of the bony structures in bilateral shoulders. Results: Patients with ASI had greater GROC in the AP direction (P < .001), HROC in the AP direction (P = .002), glenoid height (P = .005), and glenoid index (P < .001) and smaller conformity index (P < .001), glenoid width (P = .002), stability angle (P < .001), and glenoid depth (P < .001). In addition, the glenoid of the ASI patients was more anteverted compared with that of controls (P = .001). There was no statistical difference in half the measurements between the bilateral shoulder joints in patients with ASI. Conclusion: In this study, glenohumeral geometric differences were found between ASI patients and healthy control participants. Glenoid curvature and conformity index, based on bilateral comparisons of affected and contralateral shoulders, appear inherent and may predict ASI risk.

Attention Mechanism-Based Graph Neural Network Model for Effective Activity Prediction of SARS-CoV-2 Main Protease Inhibitors: Application to Drug Repurposing as Potential COVID-19 Therapy.

Compared to de novo drug discovery, drug repurposing provides a time-efficient way to treat coronavirus disease 19 (COVID-19) that is caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). SARS-CoV-2 main protease (Mpro) has been proved to be an attractive drug target due to its pivotal involvement in viral replication and transcription. Here, we present a graph neural network-based deep-learning (DL) strategy to prioritize the existing drugs for their potential therapeutic effects against SARS-CoV-2 Mpro. Mpro inhibitors were represented as molecular graphs ready for graph attention network (GAT) and graph isomorphism network (GIN) modeling for predicting the inhibitory activities. The result shows that the GAT model outperforms the GIN and other competitive models and yields satisfactory predictions for unseen Mpro inhibitors, confirming its robustness and generalization. The attention mechanism of GAT enables to capture the dominant substructures and thus to realize the interpretability of the model. Finally, we applied the optimal GAT model in conjunction with molecular docking simulations to screen the Drug Repurposing Hub (DRH) database. As a result, 18 drug hits with best consensus prediction scores and binding affinity values were identified as the potential therapeutics against COVID-19. Both the extensive literature searching and evaluations on adsorption, distribution, metabolism, excretion, and toxicity (ADMET) illustrate the premium drug-likeness and pharmacokinetic properties of the drug candidates. Overall, our work not only provides an effective GAT-based DL prediction tool for inhibitory activity of SARS-CoV-2 Mpro inhibitors but also provides theoretical guidelines for drug discovery in the COVID-19 treatment.

Efficacy of DAIR in managing Candida parapsilosis-infected total knee arthroplasty with five-year follow-up: A case report and review of literature.

RATIONALE: Fungal periprosthetic joint infections (fPJIs) are relatively uncommon, accounting for approximately 1% of all PJIs. Revision surgery is typically recommended for fungal infections; however, the physical and financial impact on patients is significant. In this report, we present a case of fPJI successfully treated with debridement, antibiotics, and implant retention (DAIR) with a favorable outcome over a 5-year period. PATIENT CONCERN: A 56-year-old male patient presented with a non-healing surgical incision 1 week after undergoing primary total knee arthroplasty on the right side. DIAGNOSIS: Microbiological culture of the wound effusion identified Candida parapsilosis. Postoperatively, the patient exhibited a significant decrease in serum albumin levels and poor glycemic control. Both C-reactive protein and erythrocyte sedimentation rate were elevated. INTERVENTIONS: A comprehensive DAIR procedure was performed, along with continuous closed irrigation using fluconazole for 1 week. The patient received intravenous voriconazole for 4 weeks, followed by oral fluconazole for an additional 3 months. OUTCOMES: At 1- and 5-year follow-up appointments, the patient C-reactive protein and erythrocyte sedimentation rate levels were within normal limits, and there was no evidence of swelling, erythema, or tenderness in the right knee joint, indicating no signs of infection. LESSONS: DAIR is an effective treatment for early fPJIs, and continuous closed irrigation may provide specific advantages. The patient nutritional status plays a crucial role in the management of periprosthetic infections.

Attention-based deep clustering method for scRNA-seq cell type identification.

Single-cell sequencing (scRNA-seq) technology provides higher resolution of cellular differences than bulk RNA sequencing and reveals the heterogeneity in biological research. The analysis of scRNA-seq datasets is premised on the subpopulation assignment. When an appropriate reference is not available, such as specific marker genes and single-cell reference atlas, unsupervised clustering approaches become the predominant option. However, the inherent sparsity and high-dimensionality of scRNA-seq datasets pose specific analytical challenges to traditional clustering methods. Therefore, a various deep learning-based methods have been proposed to address these challenges. As each method improves partially, a comprehensive method needs to be proposed. In this article, we propose a novel scRNA-seq data clustering method named AttentionAE-sc (Attention fusion AutoEncoder for single-cell). Two different scRNA-seq clustering strategies are combined through an attention mechanism, that include zero-inflated negative binomial (ZINB)-based methods dealing with the impact of dropout events and graph autoencoder (GAE)-based methods relying on information from neighbors to guide the dimension reduction. Based on an iterative fusion between denoising and topological embeddings, AttentionAE-sc can easily acquire clustering-friendly cell representations that similar cells are closer in the hidden embedding. Compared with several state-of-art baseline methods, AttentionAE-sc demonstrated excellent clustering performance on 16 real scRNA-seq datasets without the need to specify the number of groups. Additionally, AttentionAE-sc learned improved cell representations and exhibited enhanced stability and robustness. Furthermore, AttentionAE-sc achieved remarkable identification in a breast cancer single-cell atlas dataset and provided valuable insights into the heterogeneity among different cell subtypes.