Eucommia ulmoides (E. ulmoides) is the sole representative of the Eucommiaceae family, exhibiting abundant medical application as traditional Chinese medicine and high commercial application as a hardy rubber tree. Although some high-quality nuclear genomes and plastid genomes (plastome) have been reported, a comprehensive analysis of its complete mitochondrial genome (mitogenome) remains lacking. In this study, we successfully assembled the first complete mitogenome of E. ulmoides based on PacBio HiFi sequencing data. The architecture of E. ulmoides mitogenome exhibited two circular chromosomes (MC1 and MC2) with a combined length of 541.7 kb. A total of 61 genes were annotated in the E. ulmoides mitogenome, including 38 unique protein-coding genes (PCGs), 21 tRNA genes, and 2 rRNA genes. We found that the dispersed repetitive sequences comprised 21.67% of the mitogenome with the largest repeat spanning > 16 kb. The collinearity of E. ulmoides mitogenome with other species was relatively low with the longest homologous sequence being only 5,960 bp. The phylogenetic trees constructed based on mitogenomes and plastomes, respectively, consistently confirmed that E. ulmoides belongs to the basal lamiid clade, which is congruent with the APG IV classification system. The PCG composition of mitogenomes across four major lineages of land plants (liverworts, ferns, gymnosperms, and angiosperms) demonstrated the convergent reduction in mitogenomic evolution. We reported the latest mitogenome of E. ulmoides with a two-circular chromosomal architecture obtained through PacBio HiFi reads, which was the first mitogenome in the genus Eucommia and the order Garryales to be described in detail. The mitogenome of E. ulmoides provided a solid foundation for comprehending the evolutionary trajectories and elucidating the underlying causes of cytonuclear discordance in E. ulmoides.

The association between glucagon-like peptide-1 receptor agonists (GLP-1RAs) and gastrointestinal events remains controversial in observational studies. To address this, our study employed drug-target Mendelian randomization (MR) to investigate the causal relationship between GLP-1RAs and 24 gastrointestinal disorders, supplemented by sensitivity analyses, Cox proportional hazards regression analysis and mediation testing. Using single-nucleotide polymorphisms significantly associated with GLP1R gene expression as instrumental variables, we mimicked the lifelong pharmacological effects of GLP-1RAs. The significant causal association was further examined by the association study using weighted polygenic scores (PGSs) for GLP1R expression by the Cox proportional hazards regression model in UK Biobank (UKB). Our results revealed that genetically proxied GLP-1RA effects were significantly associated with a lower risk of paralytic ileus (log(OR) = −1.474, 95% CI = (−2.164, −0.784), P = 2.86 × 10 ⁻⁵ ). The direction of the association was further supported by Cox proportional hazards model using PGSs in the UKB, though the association did not reach significance ( P > .05). According to two-step mediation analysis, glycated hemoglobin A1c and body mass index only mediated 1.58% and 0.52% of the protective effect of GLP-1RAs on paralytic ileus. Unlike observational studies, our MR analysis found no causal links between GLP-1R expression and the other gastrointestinal events including cholelithiasis, pancreatitis and so on. By overcoming confounding biases inherent in observational research, this study provides genetic evidence supporting the causal protective effect of GLP-1 receptor agonists on paralytic ileus and highlights the need to reevaluate GLP-1RA-associated gastrointestinal risks in clinical contexts.

Diabetic foot is a severe chronic complication of diabetes, mainly resulting from peripheral neuropathy and vasculopathy, and may progress to ulcers, infections, and amputation if not treated in a timely manner. Early identification of patients who already present a high-risk diabetic foot profile at clinical evaluation is therefore critical for guiding preventive interventions. This study aimed to develop an interpretable machine learning-based risk classification platform for classifying current diabetic foot risk status in patients with diabetes and supporting clinical risk stratification and personalized care. The model was not designed to predict future ulcer occurrence, amputation, or wound healing outcomes. In this retrospective study, data from 1938 diabetic patients at Nanfang Hospital of Southern Medical University were used for model development, and an independent external validation cohort of 695 diabetic patients from Shanghai Changhai Hospital was used to assess generalizability. Fifty clinical features covering demographic, metabolic, vascular, neurological, inflammatory, renal, and diabetes-related factors were included. Ten machine learning models were developed and compared. Recursive Feature Elimination (RFE) was applied to the top-performing models for feature selection, and SHapley Additive exPlanations (SHAP) was used to interpret model predictions and construct a diabetic foot risk classification platform. Using sixteen selected features, the CatBoost model achieved the best performance on the internal test set, with an AUC of 0.935 ± 0.016 and accuracy, precision, and recall of 0.88. In the external validation cohort, the model maintained stable performance, with an AUC of 0.922 ± 0.006 and accuracy, precision, recall, and F1-score all equal to 0.88, demonstrating good generalizability. We developed a CatBoost-based diabetic foot risk classification model incorporating sixteen clinically accessible features. The model showed stable and reliable performance across internal and external cohorts and remained robust under data noise and class imbalance, supporting its potential utility for real-world clinical risk classification and early preventive care.

Yulania liliiflora is a highly exploitable plant with medicinal value from its flower buds, leaves, and bark. The species, as an ornamental plant, has significant implications for the phylogeny of land plants and offers promising prospects for scientific research. In this study, the Y. liliiflora mitochondrial genome (mitogenome) was assembled into a breach linear chromosome using PacBio HiFi sequencing, with a length of 865,191 bp and a GC content of 46.95%. A total of 69 genes were identified in the Y. liliiflora mitogenome. The Y. liliiflora mitogenome contains 65 unique genes, including 41 protein-coding genes, 21 tRNA genes, and three rRNA genes. Numerous repetitive sequences have been discovered, revealing the recombination events and the reason for the bigger mitogenome of Y. liliiflora. Many mitochondrial plastid sequences were found, with 12 complete chloroplast genes encompassed in these homologous fragments. Mitogenome from other Magnoliids were also used for collinearity comparison, which suggests a high recombination rate between Y. liliiflora and Magnoliids. Analysis of phylogeny has demonstrated that Y. liliiflora had the strongest genetic relation to Magnolia biondii of the Magnoliaceae, and Magnoliids emerged as a sister group to the clade encompassing monocots and eudicots. The results presented in this paper improve the understanding of the existing genetic knowledge of the genus Yulania. Simultaneously, they offer a promising avenue for conducting more comprehensive genomic investigations of Y. liliiflora.

Metabolic dysfunction-associated steatotic liver disease (MASLD) is a growing global health concern, but its early detection remains limited by the cost and accessibility of conventional diagnostic tools. This study aimed to develop an interpretable multimodal deep learning framework for MASLD classification by integrating tongue images with routine health examination data. A total of 711 matched tongue images and health records were collected from Shanghai Changhai Hospital between June and December 2024. Residual Network-50 was used to extract image features, while a graph attention network modeled structured clinical data. An attention-based fusion module integrated the 2 modalities for classification. DeepMASLD achieved an accuracy of 92.04% and an F1 score of 0.91, outperforming the corresponding unimodal models. Triglycerides, alanine aminotransferase, and total cholesterol were among the most influential structured features, while class activation maps highlighted the posterior and lateral tongue regions. This study presents an accessible, interpretable multimodal framework integrating tongue images with clinical indicators for MASLD screening, achieving 92.04% accuracy and an F1 score of 0.91. In summary, DeepMASLD may support scalable early detection, pending multicenter and histopathological validation.