Background: Understanding the genetic links between acute pancreatitis (AP) and its infectious comorbidities is crucial for prognosis and therapy, yet remains underexplored. Methods: We conducted a comprehensive post‐GWAS analysis using large‐scale summary statistics for AP and 16 infectious diseases. To pinpoint pleiotropic genes, we integrated multi‐omics data via transcriptome‐wide and proteome‐wide association studies, and resolved cell‐type‐specific effects using single‐cell analysis. Extensive locus colocalization analyses were performed to validate our findings by estimating the probability of shared causal variants. Results: This computational discovery phase prioritized 29 high‐confidence pleiotropic genes, including established loci ( SPINK1, CRP ) and novel candidates ( ERBB2, ALDH2, FLOT1 ). To functionally validate and contextualize these findings, we performed bulk transcriptomic analysis on peripheral blood from AP patients and employed gsMap, a spatial GWAS mapping algorithm, to integrate our genetic data with transcriptomics from a murine AP model, comparing pathological versus normal tissue. These analyses confirmed that the identified genes are dynamically regulated in a severity‐dependent manner in patients and are activated within specific pathological niches in pancreatic tissue. Conclusion: In conclusion, this study provides a genetic map linking AP and its infectious comorbidities, offering insights into potential prevention strategies and highlighting novel therapeutic targets for further investigation and validation.

This study developed an immune-related long non-coding RNAs (lncRNAs)-based prognostic signature by integrating multi-omics data and machine learning algorithms to predict survival and therapeutic responses in breast cancer patients. Utilizing transcriptomic and gene expression data from TCGA and GEO databases, 72 immune-related lncRNAs were identified through weighted gene co-expression network analysis (WGCNA) and ImmuLncRNA algorithms. The model was further optimized using 101 combinations of 10 machine learning approaches, ultimately constructing an immune-related lncRNA signature(IRLS) scoring system comprising nine key lncRNAs. Validated across 17 independent cohorts, the model demonstrated that high-risk patients had significantly shorter overall survival (OS) (P < 0.05), with predictive performance surpassing 95 published models (P < 0.05). Additionally, the IRLS score predicted responses to paclitaxel chemotherapy, and the low-risk group exhibited higher immune cell infiltration (P < 0.05), showing significant negative correlations with CD8A, PD-L1, tumor mutational burden (TMB), and neoantigen load (NAL). In immune checkpoint inhibitor (ICI) treatment cohorts, low IRLS scores were associated with improved response rates to atezolizumab. Our findings suggest that the IRLS model serves as a novel biomarker for prognostic stratification and personalized therapeutic decision-making in breast cancer.

The emergence of multiple variants of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) highlights the importance of possible animal-to-human (zoonotic) and human-to-animal (zooanthroponotic) transmission and potential spread within animal species. A range of animal species have been verified for SARS-CoV-2 susceptibility, either in vitro or in vivo. However, the molecular bases of such a broad host spectrum for the SARS-CoV-2 remains elusive. Here, we structurally and genetically analysed the interaction between the spike protein, with a particular focus on receptor binding domains (RBDs), of SARS-CoV-2 and its receptor angiotensin-converting enzyme 2 (ACE2) for all conceivably susceptible groups of animals to gauge the structural bases of the SARS-CoV-2 host spectrum. We describe our findings in the context of existing animal infection-based models to provide a foundation on the possible virus persistence in animals and their implications in the future eradication of COVID-19.

Cancer-associated fibroblasts (CAFs) critically influence colorectal cancer (CRC) progression and therapy response, yet their epidemiological and molecular heterogeneity remains underexplored. We integrated bulk, single-cell, and spatial transcriptomic datasets from multiple CRC cohorts, together with patient-derived tissues and functional assays, to delineate CAF subtypes and their clinical significance. Epidemiological analyses were performed across independent cohorts to evaluate the association between CAF markers and patient outcomes. A myofibroblastic CAF (myCAF) subset characterized by high LOXL2 expression was consistently enriched in advanced and chemoresistant CRC samples. Multi-omics correlation analyses revealed that LOXL2⁺ CAFs activated WNT signaling in adjacent tumor cells, promoting stemness and drug resistance. Across population-based cohorts, elevated LOXL2 expression was independently associated with poor overall and disease-free survival, as confirmed by multivariate Cox regression. Spatial transcriptomics and immunofluorescence demonstrated close physical interaction between LOXL2⁺ CAFs and WNT5A-positive cancer cells. Functional inhibition or genetic silencing of LOXL2 and wnt5a in CAFs restored chemosensitivity in vitro and suppressed tumor growth in vivo. Our integrative epidemiological and experimental analyses identify LOXL2⁺ CAFs as a key stromal determinant of chemoresistance and poor prognosis in CRC. These findings highlight a clinically relevant stromal biomarker with potential for risk stratification and therapeutic targeting in colorectal cancer.

Landslides are a type of sudden and highly destructive geological hazard. Traditional detection methods often suffer from delayed response and low efficiency. In recent years, deep learning-based object detection techniques have attracted increasing attention in disaster recognition tasks, particularly transformer-based detection models, which exhibit significant advantages in global feature modeling. However, landslide targets in remote sensing imagery often present challenges such as large-scale variation, blurred boundaries, and texture interference. To address these issues, this study proposes an improved detection algorithm based on the RT-DETR-r18 framework by integrating multiple specialized modules. First, the DDC3 module is designed to enhance the recognition of fine boundaries and local textures, thereby improving the feature extraction capacity of the backbone network. Second, an Efficient Additive Attention (EAA) mechanism is introduced to suppress redundant information and strengthen the model's focus on critical regions, improving detection precision. Finally, the CGAFusion module is employed, which utilizes a triple-attention strategy to collaboratively regulate feature weights. This module enhances the model’s ability to filter salient features while preserving global contextual information, leading to more accurate landslide edge detection. A dual-class dataset comprising landslides and storms is constructed from multi-source imagery for evaluation. The experimental results show that the proposed method outperforms existing models in several dimensions including mAP@0.5 and F1 score, demonstrating strong detection accuracy. Code is available at https://github.com/sanyauChenCoder/Landslide_02.git.

Background: Stroke remains a leading cause of morbidity and mortality in aging populations, and its risk is closely linked to metabolic disorders such as diabetes, hypertension, hyperlipidemia, and obesity. To better understand these relationships, this study aimed to quantify the associations between key metabolic health factors and both stroke incidence and cognitive outcomes in older adults, and to evaluate the predictive value of these metabolic factors for stroke risk through cross-sectional secondary analyses of two population-based cohort datasets. Methods: We analyzed data from the Health and Retirement Study (HRS; n = 7,322) and the English Longitudinal Study of Ageing (ELSA; n = 6,630). Associations with stroke incidence were assessed using multivariable logistic regression, and associations with cognitive outcomes were examined using multivariable linear regression. Random forest models evaluated the relative importance of metabolic factors for stroke prediction, with discrimination quantified by ROC curves (AUC). Mediation analyses explored whether stroke mediated the link between metabolic disorders and cognition. Results: Across both cohorts, diabetes and hypertension were consistently associated with higher odds of stroke, while obesity and hyperlipidemia showed smaller and cohort-dependent associations. Random forest analyses identified hypertension and diabetes as the strongest predictors of stroke. ROC analyses showed moderate discriminative performance for single metabolic factors (AUCs generally 0.70 – 0.80), with obesity performing weakest. Metabolic disorders—particularly diabetes and hypertension—were associated with worse cognitive performance, whereas hyperlipidemia showed small positive associations with certain cognitive measures in some models. Mediation analyses suggested that part of the adverse impact of metabolic disorders on cognition operates through stroke. Conclusions: Metabolic health is strongly linked to stroke risk and cognitive outcomes in older adults. Targeted detection and management of hypertension and diabetes should be prioritized to reduce stroke burden and cognitive decline. Given only moderate predictive performance of single metabolic markers, comprehensive risk models integrating socioeconomic, clinical, and lifestyle factors are warranted for improved stroke risk stratification.