SEARCH · PubMed Health

Results for “Random Forest”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 73 records · Page 4Linked to original sources

PubMed2025

Mapping key mitochondrial genes in Alzheimer's disease through human tissue and iPSC derived neurons.

Alzheimer's disease (AD) is a progressive neurodegenerative condition that has become a global health challenge due to an aging world population and no available effective treatment. Mitochondrial dysfunction plays a crucial role in the development of AD due to its critical role in neuronal survival and function. However, the specific mitochondrial genes and pathways involved in AD pathogenesis remain poorly defined. In this study, we incorporated seven AD human postmortem and three AD iPSC-derived neurons (iNs) gene expression datasets to identify mitochondria-related Differentially Expressed Genes (mitoDEGs) between AD and control. The Gene Ontology (GO) analysis is conducted to investigate the AD biological mechanisms, and a random forest model is developed to assess how well the key mitoDEGs differentiate AD and control groups. Through our analysis, we identified fourteen key mitochondria related genes that show significant dysregulation in both postmortem brain tissues and iNs derived from AD patients. These genes have strong connections to oxidative stress, indicating mitochondrial dysfunction plays a crucial role in Alzheimer's disease pathology. Our study identified the key genes and pathways as promising targets for future research and therapeutic interventions, highlighting the importance of mitigating oxidative stress and restoring mitochondrial function in AD.

Humans↗

PubMed2026

Extreme climatic events drive consistent and predictable shifts in soil antibiotic resistance genes.

Antimicrobial resistance (AMR) is a growing One Health challenge, and as climate warming intensifies extreme events, it remains unclear how these disturbances affect soil antibiotic resistance genes (ARGs). Here we analyzed the data from a controlled experiment using soils from 30 grassland sites across ten European countries, which simulated drought, flooding, freeze-thaw, and heatwaves to explore ARG dynamics. Overall, ARGs exhibited relatively small but highly consistent shifts across treatments. Heatwaves caused the strongest reductions in ARG abundance and in their linkages with mobile genetic elements (MGEs), a pattern that may reflect a hypothesized metabolic-genetic trade-off, in which microbial investment may shift from core metabolism toward stress signaling and structural maintenance. ARG dynamics during and after disturbance were governed by distinct soil physicochemical properties, with temperature and nutrient status determining acute responses, whereas soil moisture and seasonal variability in temperature and precipitation shaped longer-term legacy effects. Cross-validated random-forest models showed positive predictive performance for Bray-Curtis-based compositional responses within the environmental range represented by the 30 grassland sites. Our findings enhance the understanding of how soil ARGs respond to extreme climatic events and provide a step toward predicting extreme-event impacts on soil resistomes with relevance to One Health.

Soil Microbiology↗

PubMed2025

Discovery of novel diagnostic biomarkers of hepatocellular carcinoma associated with immune infiltration.

OBJECTIVE: Diagnosis of hepatocellular carcinoma (HCC) remains challenging for clinicians. Machine learning approaches and big data analyses are viable strategies for identifying HCC diagnostic markers. MATERIALS AND METHODS: In this study, we downloaded mRNA expression profiles of HCC from the GEO database and used random forest and machine learning algorithms, such as least absolute shrinkage and selection operator, to screen for reliable diagnostic genes. Disease Ontology, Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Set Enrichment Analysis enrichment analyses were performed to explore differential gene functions and disease pathways. CIBERSORT was performed to calculate the immune cell infiltration of HCC and the correlation between diagnostic genes and immune cells. Cell experiments were performed to evaluate the function of R-spondin 3 (RSPO3) in HCC cells. Immunohistochemical staining was used to evaluate the protein expression of CD138, CD206 and iNOS. RESULTS: The results indicated that extracellular matrix protein 1 (ECM1), Niemann-Pick C1-Like 1 (NPC1L1) and RSPO3 were down-regulated in HCC compared with the normal group (p < 0.05), which was validated in clinical tissue samples. Moreover, ECM1, NPC1L1 and RSPO3 had high diagnostic values (AUC > 0.75) for HCC in both training and test groups. Immuno-infiltration analysis revealed that ECM1 and RSPO3 were highly positively correlated with neutrophil and macrophage M2 levels, whereas they were negatively correlated with Tregs. RSPO3-si affected cell proliferation and apoptosis in HCC. Furthermore, RSPO3 exhibited a positive correlation with tumour progression, the proportion of plasma cells and M2 macrophages in mice, while showing a negative association with M1 macrophages. CONCLUSION: The present study identified ECM1, NPC1L1 and RSPO3 as new diagnostic biomarkers for HCC based on normal and diseased samples from HCC, meanwhile the pro-oncogenic function of RSPO3 and its regulation on immune infiltration have been confirmed.

Carcinoma, Hepatocellular↗

PubMed2026

Dynamic lysine acetylation and succinylation of platelet proteins regulates platelet storage lesion: mechanistic insights from multi-omics.

OBJECTIVES: Platelet storage lesion (PSL) severely impairs platelet function during storage, presenting a major hurdle in transfusion medicine; however, the dynamic interplay between global proteomic changes and post-translational modifications (PTMs) underlying these functional deteriorations remains insufficiently characterized. Here, we report the first comprehensive multi-omics analysis integrating global proteomics, acetylomics, and succinylomics to dissect the molecular dynamics during platelet storage. METHODS: We performed quantification of global proteomics, acetylome and succinylome based on TMT-labeled LC-MS/MS analysis, combined with antibody-affinity enrichment and purification. Dynamic molecular changes and functional transformation of platelet were also characterized under proper conditions stored for 1, 3, 5, 7 days, respectively. RESULTS: We systematically characterized 3,609 proteins, 1,308 acetylation sites, and 1,947 succinylation sites across multiple storage time points (D1, D3, D5, D7). We distinct temporal patterns of post-translational modifications, with succinylation showing more extensive coverage than acetylation in platelets. Pathway enrichment analysis revealed extensive metabolic reprogramming involving complement activation, energy metabolism, and cellular detoxification processes. The identification of specific motif patterns provided mechanistic insights into the functional specificity of these modifications. Random forest machine learning identified 20 core regulatory proteins representing critical nodes in PSL development. Furthermore, we employed real - time quantitative polymerase chain reaction (RT - QPCR) to measure the expression levels of key genes related to platelet function and PTM - associated pathways. CONCLUSION: By mapping the interplay between proteomic abundance shifts and PTM dynamics, this study provides a multidimensional understanding of PSL, establishing a foundational framework for optimizing storage protocols and enhancing transfusion safety.

Blood Platelets↗

PubMed2025

Blood-based DNA methylation markers for autism spectrum disorder identification using machine learning.

BACKGROUND: Autism spectrum disorder (ASD) is a complex neurodevelopmental disorder lacking objective biomarkers for early diagnosis. DNA methylation is a promising epigenetic marker, and machine learning offers a data-driven classification approach. However, few studies have examined whole-blood, genome-wide DNA methylation profiles for ASD diagnosis in school-aged children. METHODS: We analyzed genome-wide DNA methylation data from GEO dataset GSE113967, including 52 children with ASD and 48 typically developing (TD) controls. Differentially methylated positions (DMPs) were identified, and feature selection was performed using support vector machine-recursive feature elimination with cross-validation (SVM-RFECV). Classification models were developed using random forest (RF), extreme gradient boosting (XGBoost), and decision tree (DT) classifiers. A nomogram visualized feature contributions. RESULTS: A total of 138 DMPs differentiated ASD from TD children. Eleven CpG sites selected by SVM-RFECV formed the basis for model construction. RF and XGBoost achieved the highest accuracy (75%), with DT reaching 70%. Functional annotation indicated enrichment in cell adhesion and immune-related pathways. CONCLUSIONS: This exploratory study demonstrates the feasibility of integrating peripheral blood DNA methylation data with machine learning to distinguish children with ASD. While limited by sample size and moderate accuracy, this study provides methodological insights into the feasibility of integrating epigenetic and computational approaches for ASD-related biomarker exploration.

Humans↗

PubMed2025

MicroRNAs signatures in small extracellular vesicles for psychological resilience in young adults using machine learning.

AIMS: Psychological resilience refers to an individual's capacity to adapt to adverse events. MicroRNAs (miRNAs) play a crucial role in regulating post-transcriptional processes, while small extracellular vesicles (sEVs) act as transport vehicles. This study aimed to employ genome-wide profiling to identify and validate differences in the expression of resilience-associated sEV-miRNAs between low resilience (LR) and high resilience (HR) in young adults. METHODS: Eighty participants were divided into LR or HR based on the Connor - Davidson Resilience Scale (CD-RISC). The expression levels of the target sEV-miRNAs in LR and HR were compared and analyzed. RESULTS: Expression analyses demonstrated significant differences in let-7b, miR-151b, miR-335, and miR-193a between LR and HR (p < 0.01), with let-7b showing the highest discriminative ability. The AUC values for each sEV-miRNA ranged from 0.74 to 0.94, based on logistic regression and three machine learning models: random forest, support vector machine, and eXtreme gradient boosting. Based on leave-one-out cross-validation in different models, the combined four sEV-miRNAs demonstrated strong performance for detecting LR (AUC = 0.87-0.90). Sex-specific differences were also observed, with female participants showing more pronounced resilience signatures in targeted sEV-miRNAs. CONCLUSIONS: These findings suggest that sEV-miRNAs hold potential as biomarkers for psychological resilience in young adults.

Humans↗

PubMed2025

Integrating explainable artificial intelligence with multiomics systems biology and electronic health record data mining for personalized drug repurposing in Alzheimer's disease.

Alzheimer's disease (AD) is characterized by region- and patient-specific molecular heterogeneity, which hinders therapeutic design. In this study, we introduce PRISM-ML (PRecision-medicine using Interpretable Systems and Multiomics with Machine Learning), an open-source integrated analysis pipeline that combines interpretable machine learning with systems biology and electronic health records data mining to elucidate the molecular diversity of AD and predict promising drug repurposing opportunities. First, we integrated and harmonized transcriptomic (bulk RNA-seq) and genomic (genome-wide association study) data from 2105 brain samples, each with matched data from the same individual (1363 AD patients, 742 controls; 9 tissues), sourced from three independent studies. Random forest classifiers with SHapley Additive exPlanations identified patient-specific biomarkers; unsupervised clustering resolved 36 molecularly distinct subtissues (defined as clusters of samples within a brain tissue that share a specific expression pattern); and gene-gene coexpression networks prioritized 262 high-centrality bottleneck genes as putative regulators of dysregulated pathways. Next, knowledge graph-based drug repurposing predicted six Food and Drug Administration (FDA)-approved drugs that simultaneously target multiple bottleneck genes and multiple AD-relevant pathways. Notably, in a large US de-identified insurance-claims database (n = 364 733), exposure to promethazine, one of the candidate drugs, was associated with a 57%-62% lower incidence of AD versus an active antihistamine comparator (adjusted hazard ratio 0.38; inverse-probability weighted 0.43; both P < .001), providing real-world support for its repurposing potential. In summary, PRISM-ML, as an explainable multiomics analysis pipeline, is readily transferable to other complex diseases, advancing precision medicine.

Alzheimer Disease↗

PubMed2026

A leakage-aware genomic prediction pipeline for meropenem resistance in Klebsiella pneumoniae using transformer-based resistome representation learning.

MOTIVATION: Antimicrobial resistance (AMR) in Klebsiella pneumoniae, particularly to carbapenems such as meropenem, is a major global health problem. Machine learning is increasingly used to predict resistance from genomic markers; however, many models fail to capture high-level gene-gene interactions and may exhibit inflated performance due to lineage-biased prediction. Existing genomic prediction models largely rely on flat feature representations that fail to capture epistatic gene interactions, and commonly suffer from inflated performance estimates due to phylogenetic data leakage. To address these limitations simultaneously, a leakage-aware hybrid TabTransformer-CatBoost pipeline was developed, combining self-attention-based resistome representation learning with gradient boosting classification under clade-aware data partitioning. A self-attention encoder converts sparse gene presence-absence profiles into contextualized latent embeddings, which are subsequently classified using gradient boosting to capture lineage-aware AMR patterns. RESULTS: The proposed architecture outperformed classical baselines including Logistic Regression, Random Forest, XGBoost, and optimized CatBoost models. Internal accuracy reached 92.59% for the Chained Hybrid configuration (area under the receiver operating characteristic curve, AUROC = 0.8670, F1 = 0.8537). Performance gains primarily originated from the embedding stage, as confirmed by ablation analysis. External validation across independent multinational cohorts (n = 305) demonstrated generalizability (AUROC = 0.8105; F1 = 0.7552). Permutation testing produced near-zero Matthews Correlation Coefficient (MCC) = 0.0091, indicating predictions reflect genuine biological signal rather than noise. These results establish attention-based genomic embedding with gradient boosting as a scalable, interpretable, and leakage-aware framework for clinical AMR prediction. AVAILABILITY AND IMPLEMENTATION: The source code for the TabTransformer-CatBoost framework, including preprocessing pipelines and pre-trained embeddings, is available at https://github.com/SibelKervanci/kp-meropenem-tabtransformer.

Journal Article↗

PubMed2026

plinkQC: an integrated tool for ancestry inference, sample selection, and quality control in population genetics.

MOTIVATION: Population genetic analyses rely on high quality datasets that pass rigorous controls for sample and marker quality. Many analyses also require additional processing including identification of ancestry and sample relatedness. A software package that addresses all these common, yet crucial tasks is missing. RESULTS: We have developed plinkQC, an R/CRAN package that combines these functionalities into a single software package with detailed vignettes for example applications. plinkQC determines the ancestry of study samples via a pre-trained random forest classifier that reaches 98% performance accuracy with just 5% of marker overlap between reference and user data. To obtain the maximal set of unrelated study samples, we developed a graph-based pruning method, taking both relationship estimates and sample quality into account. We demonstrate optimal sample selection on the 1000 Genomes project, where we retain an additional 71 samples compared to publicly available exclusion lists. Finally, plinkQC bundles these results together with per-individual and per-marker quality control checks into three simple functions and returns both the quality controlled dataset and quality control report about each step of the analysis. AVAILABILITY AND IMPLEMENTATION: plinkQC is available as an R/CRAN package. The documentation and code are available on github: https://meyer-lab-cshl.github.io/plinkQC/ and https://github.com/meyer-lab-cshl/plinkQC_manuscript.

Software↗

PubMed2025

Accurate identification of abnormal ploidy using an artificial intelligence model in preimplantation genetic testing.

STUDY QUESTION: Can ultra-low-coverage whole-genome sequencing (ulc-WGS) accurately identify abnormal ploidy during preimplantation genetic testing (PGT)? SUMMARY ANSWER: The artificial intelligence (AI)-based PGT-Plus model demonstrates high accuracy in ploidy detection, offering a cost-effective solution that enhances clinical utility of PGT. WHAT IS KNOWN ALREADY: The predominant PGT for aneuploidy can identify chromosomal aneuploidies but cannot determine ploidy status. Transferring embryos with ploidy abnormalities can result in miscarriage and molar pregnancy. On the other hand, in ART, fertilization is assessed by morphological pronuclear assessment at the zygote stage. However, it has a low specificity in the prediction of abnormal ploidy status and embryos deemed abnormally fertilized can yield healthy pregnancies. Accurately identified abnormal ploidy in PGT-A can resolve current limitations and expand the utility range of PGT-A. Several studies have identified ploidy abnormalities; however, they were mainly based on single-nucleotide polymorphism (SNP) arrays or needed to combine additional targeted-next-generation sequencing (NGS) information. Studies based on ulc-WGS remain scarce. STUDY DESIGN SIZE DURATION: The study consisted of two stages: methodology establishment and validation. An AI model, named PGT-Plus, was developed using 653 samples with known ploidy status, which was further validated using 792 different ploidy status samples. In the clinical application stage, the approach was used to analyse the ploidy status of 19 103 normally fertilized PGT blastocysts and 140 single pronucleus (1PN)-derived blastocysts collected between May 2022 and December 2023. All blastocysts were tested using trophectoderm biopsy and NGS. PARTICIPANTS/MATERIALS SETTING METHODS: The methodology is based on the ulc-WGS data. First, based on samples with known ploidy status: the heterozygosity rate of high-frequency biallelic SNPs, the likelihood ratio (LLR) of alleles was calculated under different assumptions ('both parental homologs' [BPH] from a single parent, 'single parental homolog' [SPH] from each parent, disomy, and monosomy) by leveraging allele frequencies and linkage disequilibrium (LD) measured in the 1000 genomes project database. Twenty-three continuous candidate features derived from heterozygosity rates and LLRs of chromosomes or selected windows were included to establish the ploidy prediction AI model. Gini importance analysis and multicollinearity mitigation was performed for feature selection, then the performance of Random Forest (RF), Support Vector Machine (SVM), and Logistic Regression for modelling was compared. Subsequently, the parameter optimization was performed based on the RF model. Ploidy constitution concordance was evaluated in known ploidy status samples. The frequency of abnormal ploidy in normal fertilized PGT blastocysts and 1PN-derived blastocysts (including conventional IVF and ICSI) was evaluated. MAIN RESULTS AND THE ROLE OF CHANCE: Eleven features were collected for model architecture compared to SVM and Logistic Regression; RF achieved superior performance for ploidy detection. The AI model achieved an AUC of 1 for genome-wide-uniparental diploidy (GW-UPD), 1 for triploidy, and 0.99 for diploidy. For the 792 validation samples, 99.5% of samples were successfully detected using the AI model, and the model showed 100% accuracy for ploidy classification. In the clinical application stage, out of 19 103 PGT samples, 19 069 were successfully analysed using the model, with 110 (0.57%) identified as having abnormal ploidy embryos. Among these, 12.7% (14/110) were identified as GW-UPD, and 87.3% (96/110) were triploid. Among 5563 diploid blastocysts transferred, 3478 clinical pregnancies were achieved. Subsequent ploidy analysis was performed for 217 spontaneous abortion and 935 prenatal diagnostic samples, and no abnormal ploidy was identified. Furthermore, of the 140 1PN embryos tested, 40 (28.6%) exhibited GW-UPD, 3 (2.1%) exhibited triploidy, and 97 (69.3%) were determined to be biparental and normally fertilized. Among the 97 biparental embryos, 46 were diploid, 11 were mosaic, and 40 were aneuploid. In terms of the insemination pattern, the percentage of abnormal ploidy in ICSI was significantly higher than in conventional IVF (P < 0.01, 37.1% vs. 2.9%, respectively). With full informed consent, 20 patients without euploidy from normal fertilization chose 1PN-derived biparental and diploid blastocysts to transfer, resulting in 10 clinical pregnancies and 9 ongoing pregnancies. LARGE-SCALE DATA: N/A. LIMITATIONS REASONS FOR CAUTION: Some rare ploidy abnormalities, such as polyploidy with an equal number of identical sets of chromosomes and ploidy mosaicism cannot be accurately identified. Moreover, the origin of abnormal ploidy was not identified due to the unavailability of DNA from both parents. WIDER IMPLICATIONS OF THE FINDINGS: The PGT-Plus AI model provides a ploidy evaluation method based on the conventional PGT-A data and integrates directly into standard PGT-A workflows. Clinical utility results suggest that the model is a valuable tool for identifying embryos with abnormal ploidy in PGT-A and rescuing normal diploid embryos from abnormally fertilized embryos. These findings demonstrate that PGT-Plus significantly enhances the diagnostic accuracy of PGT. STUDY FUNDING/COMPETING INTERESTS: This study was supported by grants from Major Scientific Program of CITIC Group (No. 2023ZXKYB34100, to Ge.L.), Hunan Provincial Grant for Innovative Province Construction (2019SK4012), Hunan Xiangjiang New District (Changsha High-tech Zone) key core technology research project in 2023, and Science Foundation of Hunan Province (Grant 2023JJ30422). All authors declared no conflicts of interest..

artificial intelligence↗

PubMed2026

Identifying JAK2 and ANXA5 as Key Genes Linking Obstructive Sleep Apnea and Oxidative Stress via Machine Learning and Multilayer Transcriptomic Integration With Functional Validation.

Obstructive sleep apnea (OSA) is a common and severe sleep disorder closely associated with oxidative stress (OS). This study aims to identify and validate potential OS-related genes associated with OSA through bioinformatics methods. We successfully identified OS-related differentially expressed genes (OS-DEGs) by combining the limma test, weighted correlation network analysis (WGCNA), and OS-related genes from the GeneCards database. Key genes and potential biological roles were further identified using Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG), enrichment analysis, protein-protein interaction (PPI) network analysis, Lasso regression analysis, random forest algorithm, and support vector machine recursive feature elimination (SVM-RFE) method. Evaluate and validate the accuracy of key genes through receiver operating characteristic (ROC) curve analysis. The human single-cell RNA sequencing (scRNA-seq) dataset is used for cell classification annotation, analysis of key gene single-cell expression profiles, and virtual gene knockout experiments based on the scTenifoldKnk algorithm. Integrating scRNA-seq sequencing, pseudotime trajectory inference, cell-cell communication analysis, and bulk immune infiltration deconvolution reveals monocyte subtype remodeling in OSA. Finally, the expression levels of key genes in clinical samples were validated using real-time quantitative PCR (RT-qPCR) and Western blotting. A total of 57 common DEGs, indicating significant enrichment in OS, inflammation, and tumor pathways, particularly prominent in the immunometabolism pathway. By integrating DEGs, WGCNA, PPI results, and machine learning methods, key genes Janus kinase 2 (JAK2) and ANXA5 were screened out. JAK2 was significantly upregulated under disease conditions, while ANXA5 was significantly downregulated. ROC curve exhibited high accuracy (area under the curve [AUC] > 0.85). Human scRNA-seq analysis revealed that key genes were predominantly highly expressed in monocytes. Virtual knockout experiments demonstrated that these key genes play a crucial role in regulating immune responses and inflammatory reactions. PPI networks and enrichment analysis verified that downstream genes S100P, ALOX5AP, PROK2, and PADI4 may collaboratively participate in immune response and inflammation regulation. Finally, clinical sample experiment further validated the results of bioinformatics analysis. This study provides new research insights for the diagnosis, mechanism research, and treatment development of OSA in the future by integrating multilayer transcriptomic and machine learning techniques.

Humans↗

PubMed2026

Proteomic Immune Signatures of Severe HIV-Associated Tuberculosis in Sub-Saharan Africa: A Prospective, Multicenter Analysis From Uganda.

OBJECTIVES: Severe tuberculosis (TB) is a major cause of critical illness and death in people living with HIV (PLWH) worldwide. Despite this, the immunopathology of severe HIV-associated TB (HIV/TB) is poorly understood. We aimed to identify an immunopathologic signature of severe HIV/TB in sub-Saharan Africa. DESIGN AND SETTING: We analyzed proteomic data from two prospective observational cohorts of adults hospitalized with severe undifferentiated infection in Uganda: an urban discovery cohort (Entebbe, n = 241) and a rural validation cohort (Tororo, n = 253). PATIENTS: Adults (age ≥ 18 yr) hospitalized with severe febrile illness. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Across both cohorts, severe HIV/TB was common, affecting 18% of participants in the discovery cohort and 21% in the validation cohort. Overall mortality was significant (30-d mortality of 22% in the discovery cohort and 60-d mortality of 26% in the validation cohort). Participants were stratified into three HIV/TB phenotypes: HIV-negative without TB, PLWH without TB, and PLWH with microbiologically diagnosed TB. We applied ordinal random forest models in the discovery cohort as a supervised feature-selection approach to identify proteins associated with progressive HIV/TB phenotype. In both cohorts, PLWH with microbiologically diagnosed TB were at highest risk of critical illness and death (30-d mortality of 42% in the discovery cohort and 60-d mortality of 52% in the validation cohort). An eight-protein signature reliably distinguished this phenotype, reflecting mediators of macrophage/dendritic cell activation (lysosome-associated membrane glycoprotein 3), natural killer cell and T-cell stimulation and cytotoxicity (cluster of differentiation 70, class I-restricted T-cell-associated molecule), B-cell activation (immunoglobulin lambda constant 2), protease-mediated tissue injury (protease, serine 2 [trypsin-2]), dysregulated coagulation (serpin peptidase inhibitor, clade A [alpha-1 antitrypsin], member 5), extracellular matrix remodeling (epidermal growth factor-containing fibulin-like extracellular matrix protein 1), and growth hormone/insulin-like growth factor axis dysregulation (insulin-like growth factor binding protein 3). CONCLUSIONS: We identified an immunologic signature of severe HIV/TB defined by mediators of macrophage/dendritic cell and cytotoxic lymphocyte activation, extracellular matrix remodeling, and dysregulated coagulation. These findings offer new insight into HIV/TB pathobiology and highlight potential targets for host-directed therapies in this high-risk population.

Humans↗

PubMed2026

Metabolism pathway-based subtyping in pancreatic adenocarcinoma: an integrated study by bulk RNA-sequence and machine learning algorithms.

BACKGROUND: Pancreatic adenocarcinoma (PAAD) is highly aggressive, and its tumor microenvironment has significant metabolic and immune microenvironment complexity and genomic instability. In this study, by integrating the metabolic pathway activity score and clinical data, we constructed a novel risk assessment model to reveal the unique biological behavior and clinical significance behind different PAAD subtypes. METHODS: In this study, the transcriptome and clinical data of TCGA and GSE57495 databases were integrated to explore the interaction between metabolic pathways. Based on unsupervised clustering analysis of pathway activity and survival prognosis, patients with PAAD were classified into metabolic subtypes with significant prognostic differences. Subsequently, we assessed the heterogeneity of these subtypes in terms of clinical outcomes, genomic characteristics, and immune microenvironment composition. Based on the differentially expressed genes (DEGs) among metabolic subtypes, a clinical prognostic risk model and nomogram were constructed, which were double-validated by GSE57495-independent cohort and GSE57495 + TCGA-PAAD combined cohort. Finally, the correlations between risk scores (RSs) and signaling pathway activity and tumor immune microenvironment characteristics were evaluated. RESULTS: Based on metabolic pathway correlation and prognostic information, 240 patients in the TCGA-PAAD and GSE57495 datasets were divided into three subgroups. There were significant differences between subgroups in gene expression, pathway activity, clinical prognosis, and immune infiltration characteristics among the subtypes. Using machine learning algorithms, an RS model was constructed from DEGs among the subgroups, with the random forest method showing the best performance. A nomogram integrating the RS and clinical indicators demonstrated excellent predictive accuracy for 1-, 3-, and 5-year survival rates, confirming the RS as an independent prognostic factor. High- and low-risk groups exhibited significant differences in immune infiltration, pathway activity, and gene mutations. Drug sensitivity analysis showed that the high-risk group was more sensitive to AZD6244, ABT737, and other drugs. CONCLUSION: This study stratified patients with PAAD into three subgroups based on metabolic pathways and prognostic information, revealing significant differences in clinical outcomes, immune characteristics, and genetic mutations. The robust RS model developed from these findings demonstrated strong predictive power for patient survival and identified promising therapeutic strategies, providing valuable insights for advancing precision medicine in PAAD.

immune microenvironment↗

PubMed2025

Predicting Weight Loss After Vertical Sleeve Gastrectomy Using a Whole-genome Sequencing-derived Polygenic Risk Score in the All of Us Cohort.

OBJECTIVE: To create a genome-wide polygenic risk score (PRS) to improve prediction of a 12-month percentage weight loss (WL) after vertical sleeve gastrectomy (VSG). BACKGROUND: Variability in post-VSG WL is not well explained by clinical factors. The All of Us program provides access to a 414,830 short-read whole-genome sequencing resource, enabling unbiased discovery of genetic predictors after VSG. METHODS: VSG counts, demographic, anthropomorphic and vital sign information were obtained from the linked electronic health record. The discovery cohort (DC) included participants from version 7 carried into version 8 while the validation cohort (VC) included those newly added to v8. We defined good responders and nonresponders as having WL±1SD from the mean. Following quality filtering, we applied a 2-stage penalized-regression, followed by elastic-net logistic regression, to identify 1583 stable variants and derive β-weights. We then tested this PRS on the DC into a prediction model. RESULTS: We identified 395 participants in the DC and 336 participants in the VC, respectively. Of these, VSG, 44 were classified as good responders (≥37% WL) and 55 as nonresponders (≤19% WL). In the VC, 55 were classified as good responders and 48 as nonresponders. Adding the PRS to models to clinical predictors increased the area under the curve following logistic regression by 0.03; P <4.3 × 10 -14 , random forest by 0.03; P <9.1 × 10 -7 , decision tree by 0.05; P = 1.2 × 10 -3 , and gradient boosting by 0.08; P <8.3 × 10 -10 . CONCLUSIONS: Use of short-read whole-genome sequencing from All of Us (AoU) can be effectively used to generate PRS to enhance predictive WL accuracy. This work has implications for outcomes of both bariatric surgery and other surgical procedures.

Humans↗

PubMed2023

Evaluating the impact of modeling choices on the performance of integrated genetic and clinical models.

The value of genetic information for improving the performance of clinical risk prediction models has yielded variable conclusions. Many methodological decisions have the potential to contribute to differential results across studies. Here, we performed multiple modeling experiments integrating clinical and demographic data from electronic health records (EHR) and genetic data to understand which decision points may affect performance. Clinical data in the form of structured diagnostic codes, medications, procedural codes, and demographics were extracted from two large independent health systems and polygenic risk scores (PRS) were generated across all patients with genetic data in the corresponding biobanks. Crohn's disease was used as the model phenotype based on its substantial genetic component, established EHR-based definition, and sufficient prevalence for model training and testing. We investigated the impact of PRS integration method, as well as choices regarding training sample, model complexity, and performance metrics. Overall, our results show that including PRS resulted in higher performance by some metrics but the gain in performance was only robust when combined with demographic data alone. Improvements were inconsistent or negligible after including additional clinical information. The impact of genetic information on performance also varied by PRS integration method, with a small improvement in some cases from combining PRS with the output of a clinical model (late-fusion) compared to its inclusion an additional feature (early-fusion). The effects of other modeling decisions varied between institutions though performance increased with more compute-intensive models such as random forest. This work highlights the importance of considering methodological decision points in interpreting the impact on prediction performance when including PRS information in clinical models.

Preprint↗

PubMed2025

Demographics, Overlap, and Latency of Severe Cutaneous Adverse Reactions in an FDA Database.

IMPORTANCE: Severe cutaneous adverse reactions (SCARs), including Stevens-Johnson syndrome/toxic epidermal necrolysis (SJS-TEN), drug reaction with eosinophilia and systemic symptoms (DRESS), acute generalized exanthematous pustulosis (AGEP), and generalized bullous fixed drug eruption (GBFDE), are rare but life-threatening drug hypersensitivity syndromes. Due to their low incidence and diagnostic complexity, large-scale characterization of SCAR is challenging. OBJECTIVE: To characterize the demographics, causative agents, trends, latency, and phenotypic overlap of SCAR using a large-scale, sanitized pharmacovigilance dataset from FAERS (FDA Adverse Event Reporting System). DESIGN: Cross-sectional study of spontaneous adverse event reports. Cases were drawn from the U.S. Food and Drug Administration Adverse Event Reporting System (FDA FAERS) from January 2004 to December 2023 and subjected to sanitization and deduplication. Disproportionality analysis was used to characterize causative agents. Machine learning (random forest classifiers) was used to analyze predictors of drug latency and mortality. SETTING: Global pharmacovigilance reports submitted to FAERS. PARTICIPANTS: A total of 56,683 deduplicated SCAR reports were identified, representing 0.33% of reports during the study period. EXPOSURES: Suspected causative drugs, including both small molecules and biologics. MAIN OUTCOMES AND MEASURES: Main outcomes included the frequency and distribution of SCAR syndromes, reporting trends over time, latency from drug start to reaction onset, drug-specific disproportionality (PRR, ROR, IC), and co-reporting between SCAR types and related conditions. RESULTS: A total of 56,683 unique SCAR reports were identified, including SJS-TEN (28,871), DRESS (22,444), AGEP (6,183), and GBFDE (150). We identified 237 drugs with significant disproportionality for SCAR overall. Co-reporting between SCARs was significantly enriched (p < 1e-200), suggesting overlapping phenotypes. Latency varied by drug and syndrome (median: GBFDE 3 days, AGEP 4 days, SJS-TEN 12 days, DRESS 20 days). CONCLUSIONS AND RELEVANCE: SCAR syndromes display distinct but overlapping phenotypes, with variable latency and diverse causative agents. These findings, based on the largest SCAR dataset to date, highlight the need for improved classification frameworks and molecular validation. Large-scale pharmacovigilance, integrated with genomic and histopathologic data, will be critical to improving diagnosis, mechanistic understanding, and clinical management of SCAR.

Acute Generalized Exanthematous Pustulosis↗

PubMed2025

Integrating explainable AI with multiomics systems biology and EHR data mining for personalized drug repurposing in Alzheimer's disease.

Alzheimer's disease (AD) is characterized by region- and patient-specific molecular heterogeneity, which hinders therapeutic design. In this study, we introduce PRISM-ML (PRecision-medicine using Interpretable Systems and Multiomics with Machine Learning), an open-source integrated analysis pipeline that combines interpretable machine learning with systems biology and electronic health record (EHR) data mining to elucidate the molecular diversity of AD and predict promising drug repurposing opportunities. First, we integrated and harmonized transcriptomic (bulk RNA-seq) and genomic (genome-wide association study) data from 2105 brain samples, each with matched data from the same individual (1363 AD patients, 742 controls; nine tissues), sourced from three independent studies. Random forest classifiers with SHapley Additive exPlanations (SHAP) identified patient-specific biomarkers; unsupervised clustering resolved 36 molecularly distinct "subtissues" (clusters of samples); and gene-gene co-expression networks prioritized 262 high-centrality bottleneck genes as putative regulators of dysregulated pathways. Next, knowledge graph-based drug repurposing predicted six FDA-approved drugs that simultaneously target multiple bottleneck genes and multiple AD-relevant pathways. Notably, in a large U.S. de-identified insurance-claims database (n = 364733), exposure to promethazine, one of the candidate drugs, was associated with a 57-62 % lower incidence of AD versus an active antihistamine comparator (adjusted hazard ratio 0.38; inverse-probability weighted 0.43; both p < 0.001), providing real-world support for its repurposing potential. In summary, PRISM-ML, as an explainable multi-omics analysis pipeline, is readily transferable to other complex diseases, advancing precision medicine.

Computational Biology↗

PubMed2025

Machine Learning-Based Identification of Survival-Associated CpG Biomarkers in Pancreatic Ductal Adenocarcinoma.

Pancreatic ductal adenocarcinoma (PDAC) is an exceptionally aggressive cancer with a 5-year survival rate of less than 10%, driven by late-stage diagnosis, limited treatment options, and a lack of reliable biomarkers for early detection and prognosis. In this study, we integrated DNA methylation data from TCGA and ICGC cohorts, categorizing samples based on survival time, and identified 684 differentially methylated CpG sites, along with 224 CpG biomarkers significantly associated with patient survival through statistical and machine learning-based analyses. We developed a random forest model to predict patient survival, achieving 85.2% accuracy for short-survival patients and 70.0% for long-survival patients in the validation set. External dataset validation further confirmed the model's robustness and accuracy. De novo motif analysis of genomic regions surrounding the 224 CpG biomarkers identified TWIST1 and FOXA2 as key transcriptional regulators enriched in survival-associated CpG sites, linking their activity to patient survival outcomes. Collectively, our findings highlight valuable epigenetic biomarkers and provide a predictive model to assess PDAC risk levels post-surgery, offering the potential for improved patient stratification and personalized therapeutic strategies.

Journal Article↗