Macrovascular complications are leading causes of morbidity and mortality in patients with type 2 diabetes mellitus (T2DM), yet early diagnosis of cardiovascular disease (CVD) in this population remains clinically challenging. This study aims to develop a machine learning model that can accurately predict diabetic macroangiopathy in Chinese patients. A retrospective cross-sectional analytical study was conducted on 1566 hospitalized patients with T2DM. Feature selection was performed using recursive feature elimination (RFE) within the mlr3 framework. Model performance was benchmarked using 29 machine learning (ML) models, with the ranger model selected for its superior performance. Hyperparameters were optimized through grid search and 5-fold cross-validation. Model interpretability was enhanced using SHAP values and PDPs. An external validation set of 106 patients was used to test the model. Key predictive variables identified included the duration of T2DM, age, fibrinogen, and serum urea nitrogen. The predictive model for macroangiopathy was established and showed good discrimination performance with an accuracy of 0.716 and an AUC of 0.777 in the training set. Validation on the external dataset confirmed its robustness with an AUC of 0.745. This study establish an approach based on machine learning algorithm in features selection and the development of prediction tools for diabetic macroangiopathy.

Background: Selecting features related to postoperative infection following cardiac surgery was highly valuable for effective intervention. We used machine learning methods to identify critical perioperative infection-related variables after mitral valve surgery and construct a prediction model.Methods: Participants comprised 1223 patients who underwent cardiac valvular surgery at eight large centers in China. The ninety-one demographic and perioperative parameters were collected. Random forest (RF) and least absolute shrinkage and selection operator (LASSO) techniques were used to identify postoperative infection-related variables; the Venn diagram determined overlapping variables. The following ML methods: random forest (RF), extreme gradient boosting (XGBoost), Support Vector Machine (SVM), Gradient Boosting Decision Tree (GBDT), AdaBoost, Naive Bayesian (NB), Logistic Regression (LogicR), Neural Networks (nnet) and artificial neural network (ANN) were developed to construct the models. We constructed receiver operating characteristic (ROC) curves and the area under the ROC curve (AUC) was calculated to evaluate model performance.Results: We identified 47 and 35 variables with RF and LASSO, respectively. Twenty-one overlapping variables were finally selected for model construction: age, weight, hospital stay, total red blood cell (RBC) and total fresh frozen plasma (FFP) transfusions, New York Heart Association (NYHA) class, preoperative creatinine, left ventricular ejection fraction (LVEF), RBC count, platelet (PLT) count, prothrombin time, intraoperative autologous blood, total output, total input, aortic cross-clamp (ACC) time, postoperative white blood cell (WBC) count, aspartate aminotransferase (AST), alanine aminotransferase (ALT), PLT count, hemoglobin (Hb), and LVEF. The prediction models for infection after mitral valve surgery were established based on these variables, and they all showed excellent discrimination performance in the test set (AUC > 0.79).Conclusions: Key features selected by machine learning methods can accurately predict infection after mitral valve surgery, guiding physicians in taking appropriate preventive measures and diminishing the infection risk.

Lung adenocarcinoma (LUAD) is the most common type of lung cancer and the leading cause of cancer incidence and mortality worldwide. Despite the improvement of traditional and immunological therapies, the clinical outcome of LUAD is still far from satisfactory. Patients given the same treatment regimen had different responses and clinical outcomes due to the heterogeneity of LUAD. How to identify the targets based on heterogeneity analysis is crucial for treatment strategies. Recently, the single-cell RNA-sequencing (scRNA-seq) technology has been used to investigate the tumor microenvironment (TME) based on cell-specific changes and shows prominently valuable for biomarker prediction. In this study, we systematically analyzed a meta-dataset from the multiple LUAD scRNA-seq datasets in LUAD, identified 15 main types of cells and 57 cell subgroups, and revealed a series of potential biomarkers in M2b, exhausted CD8⁺T, endothelial cells, fibroblast, and metabolic patterns in TME, which further validated with immunofluorescence in clinical cohorts of LUAD. In the prognosis analysis, M0 macrophage and T cell activation were shown correlated to a better prognosis (p<0.05). Briefly, our study provided insights into the heterogeneity of LUAD and assisted in novel therapeutic strategies for clinical outcome improvement.

Background: Glioma is the most common intracranial malignancy with a poor prognosis. Although remarkable advances have been made in the study of diagnostic and prognostic biomarkers, the efficacy of current treatment strategies is still unsatisfactory. Therefore, developing novel and reliable targets is desperately needed for glioma patients. Pyroptosis reshapes tumor immune microenvironment (TME) and promotes the destruction of the tumor by the immune system. Moreover, pyroptosis levels correlate with prognosis and immunotherapy response in many cancer patients. This study performed a comprehensive analysis of pyroptosis in the glioma, unveiling its potential value in glioma prognosis prediction and therapy efficacy. Methods: Firstly, the pyroptosis regulation patterns were comprehensively evaluated on 33 pyroptosis-related genes in 1716 glioma samples. The correlations were analyzed between pyroptosis regulation patterns and TME immune cell infiltration properties. Next, pyroptosis regulation patterns were measured by the PSscore model based on principal component analysis algorithms. The correlations were analyzed between PSscore and tumor mutational burden (TMB), immune checkpoint blockade (ICB) therapeutic advantages. Last, the findings were validated in an independently collected external clinical cohort. Results: We determined two distinct pyroptosis regulation patterns. The cluster-A was high immune cell infiltration with a poor prognosis (p < 0.001), whereas the cluster-B was low immune cell infiltration with a better prognosis (p < 0.001). We developed the PSscore as a measure for pyroptosis regulation patterns. The high PSscore with an inflamed TME phenotype, a high TMB (p < 0.0001), increased innate immune response, and a poor prognosis (p < 0.001). It was in stark contrast to the low PSscore (p < 0.001). Analysis of PSscore with checkpoint therapy indicated high PSscore were correlated with enhanced response to anti-PD-1 immunotherapy (p = 0.0046). For validation, we utilized in vitro experiments on an external clinical cohort. The results demonstrated that GSDMD expression level in the high PSscore group was significantly upregulated compared to the low PSscore group (p < 0.001); the CD3+ T cells and the CD3+PD-1+ cells significantly increased in the high PSscore group compared to the low PSscore group (p < 0.01). Conclusions: The PSscore of pyroptosis regulation pattern is a reliable biomarker, and it is valuable to predict prognosis, TME, and ICB therapeutic efficiency in glioma patients.

Background Prognostic indicators in lung adenocarcinoma (LUAD) have been seeking under database analysis, and remarkable advance is on the way. Methods This study calculated the scores of stromal and immune components of the tumor microenvironment (TME) in 551 LUAD samples using the ESTIMATE algorithm on The Cancer Genome Atlas (TCGA) database. R package ''limma'' was used to selected differentially expressed genes (DEG). We have analyzed the DEGs by means of Gene Ontology (GO) analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichments. The protein‐protein network, univariate Cox analysis, and Lasso regression analysis were performed to selected survival‐related genes. Gene Set Enrichment Analysis (GSEA) represented the enriched pathway of CC chemokine receptor 2 (CCR2). The ratios of immune cells in the TME of each LUAD sample were obtained using the R package "limma" and CIBERSORT algorithm in R 4.0.2. Results The ImmuneScore was positively correlated with prognosis regarding survival rate, T classification of TNM stages, and clinicopathological staging characteristics. GO and KEGG enrichments showed DEGs were associated with immune‐related activities. Three genes of LUAD were selected from the PPI network and Cox proportional hazards regression analysis. CCR2 was the most survival correlated gene by Lasso regression analysis. GSEA results showed that C2 kegg gene sets in the CCR2 high‐expression group were mainly enriched in the B cell or T cell receptor signaling pathway and natural killer cell‐mediated cytotoxicity. Correlation of CCR2 expression with prognosis was conducted, implicating a positive correlation with the prognosis of survival rate and M classification, negative correlation with the prognosis of T and N classifications. The correlation between CCR2 and tumor‐infiltrating immune cells (TICs) was analyzed, and 14 kinds of TICs were found closely correlated with CCR2 expression through difference analysis. Conclusion Therefore, CCR2 has prognostic value as an immune indicator in LUAD.

The obligate intracellular bacterium Chlamydia trachomatis is an important human pathogen whose biphasic developmental cycle consists of an infectious elementary body and a replicative reticulate body. Whereas σ ⁶⁶ , the primary sigma factor, is necessary for transcription of most chlamydial genes throughout the developmental cycle, σ ²⁸ is required for expression of some late genes. We previously showed that the Chlamydia- specific transcription factor GrgA physically interacts with both of these sigma factors and activates transcription from σ ⁶⁶ - and σ ²⁸ -dependent promoters in vitro . Here, we investigate the organismal functions of GrgA. We show that GrgA overexpression decreased RB proliferation via time-dependent transcriptomic changes. Significantly, σ ⁶⁶ -dependent genes that code for two important transcription repressors are among the direct targets of GrgA. One of these repressors is Euo, which prevents the expression of late genes during early phases. The other is HrcA, which regulates gene expression in response to heat shock. The direct regulon of GrgA also includes a σ ²⁸ -dependent gene that codes for the putative virulence factor PmpI. Conditional overexpression of Euo and HrcA also inhibited chlamydial growth and affected GrgA expression. Transcriptomic studies suggest that GrgA, Euo, and HrcA have distinct but overlapping indirect regulons. Furthermore, overexpression of either GrgA leads to decreased expression of numerous tRNAs. These findings indicate that a GrgA-mediated transcriptional regulatory network controls C. trachomatis growth and development. IMPORTANCE: Chlamydia trachomatis is the most prevalent sexually transmitted bacterial pathogen worldwide and is a leading cause of preventable blindness in under-developed areas as well as developed countries. Previous studies showed that the novel transcription factor GrgA activated chlamydial gene transcription in vitro , but did not addressed the organismal function of GrgA. Here, we demonstrate growth inhibition in C. trachomatis engineered to conditionally overexpress GrgA. GrgA overexpression immediately increases the expression of two other critical transcription factors (Euo and HrcA) and a candidate virulence factor (PmpI), among several other genes. We also reveal chlamydial growth reduction and transcriptomic changes including decreased GrgA mRNA levels in response to either Euo or HrcA overexpression. Thus, the transcription network controlled by GrgA likely plays a crucial role in chlamydial growth and pathogenesis.