Zheng Ye

Antimicrobial resistance (AMR) is a global health crisis shaped by complex ecological and evolutionary processes that often occur in polymicrobial communities. Metagenomics enables culture-independent profiling of microbial DNA directly from clinical or environmental samples, providing an unparalleled view of community composition, resistome content, and the mobile genetic elements that drive horizontal gene transfer (HGT). Yet, a recurring challenge is that metagenomic detection of antibiotic-resistance genes does not automatically translate into a mechanistic understanding of resistance phenotypes, nor does it replace culture-based functional validation. Here, we synthesize how modern metagenomics supports AMR research across three linked questions: (i) what resistance determinants are present and how do they change across time and space, (ii) which hosts and mobile genetic elements carry these determinants, and how gene flow can be inferred, and (iii) what evidence is required to move from “resistance potential” to robust mechanistic claims. We emphasize practical design principles (sampling, controls, and contamination management), analytical choices (database and parameter effects), and recent advances, including long-read sequencing for resolving antibiotic-resistance genes context, and rapid clinical metagenomic sequencing for time-sensitive decision support. We propose an evidence ladder for mechanistic inference that integrates metagenomics with targeted assays and culture-dependent experiments. Beyond synthesizing recent advances, this review provides operational tools for critical appraisal and study design: an evidence ladder for mechanistic inference, a decision-gated workflow that ties metagenomic outputs to allowable claim language, a minimum reporting checklist aligned to evidence strength, and a “pitfall → consequence → fix” guide to reduce over-interpretation. To support a more comprehensive, forward-looking view, we also summarize emerging directions that are rapidly reshaping AMR metagenomics—multi-omics integration, single-cell, and epigenetic linkage strategies, CRISPR-enabled enrichment/depletion, and AI-assisted discovery/mining—and clarify where these advances strengthen (or do not strengthen) mechanistic claims within the same evidence ladder.

Head and neck squamous cell carcinoma (HNSCC) rank among the most prevalent types of head and neck cancer globally. Unfortunately, a significant number of patients receive their diagnoses at advanced stages, limiting the effectiveness of available treatments. The tumor microenvironment (TME) is a pivotal player in HNSCC development, with macrophages holding a central role. Macrophages demonstrate diverse functions within the TME, both inhibiting and facilitating cancer progression. M1 macrophages are characterized by their phagocytic and immune activities, while M2 macrophages tend to promote inflammation and immunosuppression. Striking a balance between these different polarization states is essential for maintaining overall health, yet in the context of tumors, M2 macrophages typically prevail. Recent efforts have been directed at controlling the polarization states of macrophages, paving the way for novel approaches to cancer treatment. Various drugs and immunotherapies, including innovative treatments based on macrophages like engineering macrophages and CAR-M cell therapy, have been developed. This article provides an overview of the roles played by macrophages in HNSCC, explores potential therapeutic targets and strategies, and presents fresh perspectives on the future of HNSCC treatment.

Introduction: Natural menopause is an inevitable biological process with significant implications for women's health. However, the molecular mechanisms underlying menopause are not well understood. This study aimed to investigate the molecular and cellular changes occurring in the ovary before and after perimenopause.Methods: Single-cell sequencing data from the GTEx V8 cohort (30-39: 14 individuals; 40-49: 37 individuals; 50-59: 61 individuals) and transcriptome sequencing data from ovarian tissue were analyzed. Seurat was used for single-cell sequencing data analysis, while harmony was employed for data integration. Cell differentiation trajectories were inferred using CytoTrace. CIBERSORTX assessed cell infiltration scores in ovarian tissue. WGCNA evaluated co-expression network characteristics in pre- and post-perimenopausal ovarian tissue. Functional enrichment analysis of co-expression modules was conducted using ClusterprofileR and Metascape. DESeq2 performed differential expression analysis. Master regulator analysis and signaling pathway activity analysis were carried out using MsViper and Progeny, respectively. Machine learning models were constructed using Orange3.Results: We identified the differentiation trajectory of follicular cells in the ovary as ARID5B+ Granulosa -> JUN+ Granulosa -> KRT18+ Granulosa -> MT-CO2+ Granulosa -> GSTA1+ Granulosa -> HMGB1+ Granulosa. Genes driving Granulosa differentiation, including RBP1, TMSB10, SERPINE2, and TMSB4X, were enriched in ATP-dependent activity regulation pathways. Genes involved in maintaining the Granulosa state, such as DCN, ARID5B, EIF1, and HSP90AB1, were enriched in the response to unfolded protein and chaperone-mediated protein complex assembly pathways. Increased contents of terminally differentiated HMGB1+ Granulosa and GSTA1+ Granulosa were observed in the ovaries of individuals aged 50-69. Signaling pathway activity analysis indicated a gradual decrease in TGFb and MAPK pathway activity with menopause progression, while p53 pathway activity increased. Master regulator analysis revealed significant activation of transcription factors FOXR1, OTX2, MYBL2, HNF1A, and FOXN4 in the 30-39 age group, and GLI1, SMAD1, SMAD7, APP, and EGR1 in the 40-49 age group. Additionally, a diagnostic model based on 16 transcription factors (Logistic Regression L2) achieved reliable performance in determining ovarian status before and after perimenopause.Conclusion: This study provides insights into the molecular and cellular mechanisms underlying natural menopause in the ovary. The findings contribute to our understanding of perimenopausal changes and offer a foundation for health management strategies for women during this transition.

Head and neck cancer is a highly heterogeneous malignant tumor. Numerous studies have shown that the immune microenvironment of head and neck cancer has a significant impact on its occurrence and development, as well as its prognosis. However, there have been fewer studies related to the accurate immunophenotyping of head and neck cancer. In this study, we used gene expression profile information and clinical information from the TCGA-HNSC cohort (502 samples) and the GSE655858 cohort (270 samples) to identify and independently validate three immune subtypes (Cluster1–Cluster3) with different immune-related molecular profiles and clinical outcomes. Cluster2, which is mainly dominated by B-lymphocyte infiltration, was found to have the best prognosis. In addition, a support vector machine (SVM)-based classifier was constructed, which could accurately classify HNSC based on 19 genes. Furthermore, the results of the prognostic analysis showed activation of antibody-secreting B-lymphocyte function, which showed a good prognostic effect in all three immune subtypes of HNSC. Finally, the immune evolutionary landscape of HNSC was constructed in an attempt to explain the evolutionary pattern of the immune subtypes of HNSC. In summary, we provide a conceptual framework for understanding the tumor immune microenvironment in HNSC and demonstrate the importance of immune infiltration of B lymphocytes in HNSC. Further research is needed to assess the importance of these immunophenotypes in combination drug therapy and to provide a basis for screening appropriate patients for immunotherapy.

Post-acute pancreatitis diabetes mellitus (PPDM-A) is the main component of pancreatic exocrine diabetes mellitus. Timely diagnosis of PPDM-A improves patient outcomes and the mitigation of burdens and costs. We aimed to determine risk factors prospectively and predictors of PPDM-A in China, focusing on giving personalized treatment recommendations. Here, we identify and evaluate the best set of predictors of PPDM-A prospectively using retrospective data from 820 patients with acute pancreatitis at four centers by machine learning approaches. We used the L1 regularized logistic regression model to diagnose early PPDM-A via nine clinical variables identified as the best predictors. The model performed well, obtaining the best AUC = 0.819 and F1 = 0.357 in the test set. We interpreted and personalized the model through nomograms and Shapley values. Our model can accurately predict the occurrence of PPDM-A based on just nine clinical pieces of information and allows for early intervention in potential PPDM-A patients through personalized analysis. Future retrospective and prospective studies with multicentre, large sample populations are needed to assess the actual clinical value of the model.