Monkeypox virus (Mpox) has recently drawn global attention due to outbreaks beyond its traditional endemic regions. Understanding the immune response to Mpox infection is essential for improving disease management and guiding vaccine development. In this study, we used several machine learning algorithms to analyze time series gene expression data from macaques infected with Mpox, aiming to uncover key immune-related genes involved in different stages of infection. The dataset covered early infection, late infection, and rechallenge phases. We applied nine feature ranking methods to analyze the feature importance, obtaining nine feature lists. Then, the incremental feature selection method was applied to each list to extract key genes and build efficient prediction models and classification rules for each list. This procedure employed twelve classification algorithms and the Synthetic Minority Oversampling Technique. Key genes—such as CD19, MS4A1, and TLR10—were repeatedly identified from multiple feature lists, and are known to play vital roles in B-cell activation, antibody production, and innate immunity. Furthermore, we identified several novel key genes (HS3ST1, SPAG16, and MTARC2) that have not been reported previously. These findings offer valuable insights into the host immune response and highlight potential molecular targets for monitoring and intervention in Mpox infections.

Coronavirus Disease 2019 (COVID-19) and other lower respiratory tract diseases (LRTDs), including bacterial pneumonia and acute respiratory distress syndrome, share overlapping clinical features but arise from distinct pathophysiological mechanisms. The molecular signatures that distinguish these diseases remain insufficiently characterized in African populations, where genetic background, endemic infections, and environmental exposures may substantially shape immune responses. We integrated spatially resolved single-cell transcriptomic profiles from lung autopsy specimens of 30 Malawian patients, including 10 with COVID-19, 12 with other LRTDs, and 8 non-LRTD controls. In total, 61,391 cells representing 15 cell types and 36,602 gene expression features were analyzed. Using an integrated machine learning framework that combined nine feature-ranking algorithms with incremental feature selection, we identified potential molecular signatures that could discriminate among disease states within this cohort. The optimal classification models achieved weighted F1 scores greater than 0.94, demonstrating a robust capacity to differentiate COVID-19 from other LRTDs in our dataset. Notably, the macrophage-associated state in COVID-19 was dominated by an IFN-γ response with upregulation of CD163 and HLA-DQA2, contrasting sharply with the type I/III interferon signature reported in European cohorts. In addition, we observed cell-type-specific COVID-19 signatures, including downregulation of CAV1 in AT1 cells, consistent with epithelial damage; dysregulation of SFTPC in AT2 cells, suggesting surfactant dysfunction; and upregulation of NFKBIA in neutrophils, indicating altered inflammatory regulation. Gene Ontology enrichment further revealed universal disruption of protein synthesis machinery, along with cell-type-specific alterations in immune activation, epithelial repair, and inflammatory signaling pathways.

Choroidal neovascularization (CNV) is a common and serious complication in various retinal diseases that involves the growth of new blood vessels from the choroid into the subretinal space, leading to a visual threat. Its underlying mechanism has not been fully uncovered. Discovering new genes related to CNV is an important way to reveal its molecular mechanisms. In this study, we used DisGeNET as the initial data source to identify genes related to CNV. The gene cluster method and additional screening tests were designed to explore the genetic landscape related to CNV. Our comprehensive analysis revealed many genes with high confidence, specifically identifying novel candidates that traditional topological methods missed. Notably, we identified SERPINE1, COL1A1, and ANGPT1 as unique gene signatures using our network embedding approach. Functionally, SERPINE1 regulates the plasminogen activation system to control proteolytic balance, COL1A1 maintains the structural integrity of the extracellular matrix during invasion, and ANGPT1 is critical for the maturation and stabilization of nascent vessels. Meanwhile, some genes were also identified by other methods, such as MMP2, a regulator of extracellular matrix degradation. The newly found genes are essential for understanding the process that initiates CNV.

MicroRNA (miRNA) regulation plays a pivotal role in intracellular gene expression. Analysis of miRNA profiles can provide critical insights into disease states. As cancer-associated molecules reported in previous studies, miRNAs may serve as candidate classificatory features for exploratory cancer classification. This research analyzed serum miRNA data from patients with 13 solid cancer types and individuals without cancer. The study comprised two distinct analyses: first, stratifying the dataset into cancer and non-cancer groups to identify miRNAs differentially represented in cancer patients; and second, subdividing the cancer patient data into 13 predefined solid-cancer types to identify candidate miRNA features that discriminate among these cancer types. We employed seven feature-ranking algorithms to evaluate miRNA contributions in both analyses and generate feature lists. Each list was examined using an incremental feature selection method to extract essential miRNAs and build good-performing classification models. Several candidate miRNAs were identified for distinguishing pan-cancer samples from non-cancer ones: miR-4783-3p has been linked to associated with the regulation of endocrine cell differentiation, and miR-663a has been reported in hepatocellular carcinoma and thyroid carcinoma. The analysis also highlighted miRNAs that differentiate solid cancer types, including miR-629-3p, reported to be upregulated in lung and breast cancer, and miR-6087, reported to be downregulated in osteosarcoma and bladder cancer.

Corneal transplantation (CT) is the primary treatment for blindness caused by corneal disease, but graft rejection remains a critical determinant of long-term prognosis. Conventional rejection assessments are typically retrospective and may miss opportunities for early intervention. This study used machine-learning algorithms to examine peripheral blood gene-expression profiles in corneal transplant recipients before and after transplantation. Publicly available transcriptomic data comprised 240 pre-transplant samples (102 high-risk and 138 low-risk) and 332 post-transplant samples (193 high-risk and 139 low-risk). Each sample was represented by expression measurements for 20,522 genes. Nine feature-selection methods, two classification algorithms, incremental feature selection, and the Synthetic Minority Over-sampling Technique were used to analyze the before-CT and after-CT datasets separately for biomarker screening. Intersection analysis of the optimal feature subsets identified ABCD1, LOC100129307, SF3B4, and RAMP2 as important genes in both the before-CT and after-CT settings. Risk groups were defined clinically before transplantation and were not based on rejection events; acute rejection was not analyzed as an outcome. These findings constitute exploratory feature discovery and identify candidate peripheral blood transcriptomic features associated with clinical risk status for further study; they do not establish independently validated biomarkers or generalizable predictive performance.

In the context of chronic obstructive pulmonary disease (COPD), understanding gene expression patterns across various cell types is pivotal for disease comprehension, therapeutic development, and enhancement of patient prognosis. Nevertheless, the intricate interplay between immune and structural cells, along with their significance in COPD pathogenesis, remains only partially understood. This research employed multiple machine learning techniques to pinpoint marker genes exhibiting pronounced alterations between COPD patients and healthy individuals. We utilized six feature selection algorithms (LASSO, LightGBM, MCFS, RF, CATboost, and XGBoost) and three classification methods (DT, KNN, and RF) to analyze single-cell patient data. Cells were bifurcated into two subsets: immune cells and epithelial–stromal cells. Each cell, encompassing 21,237 genes, was labeled either as a healthy control or COPD. We identified a number of promising genes, such as FKBP5, which encodes a protein crucial for stress response regulation, and SCGB3A1, which is predominantly expressed in the respiratory tract and critical for regulating lung function. Both genes exhibit strong correlations with COPD. The study also underscored the potential roles of HLA-DRB1 and HSP90AA1 in COPD pathogenesis. Our findings underscore the unparalleled potential of machine learning in identifying genetic markers for intricate diseases like COPD, paving the way for targeted therapeutic advancements.