SEARCH · PubMed Health

Results for “Machine Learning”

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 199 records · Page 11Linked to original sources

PubMed2026

ProMeta: a meta-learning framework for robust disease diagnosis and prediction from plasma proteomics.

MOTIVATION: The plasma proteome offers a dynamic window of human health, capturing the real-time intersections between genetics and physiology. However, the application of deep learning to proteomics is currently hindered by a reliance on large-scale labeled datasets, rendering standard models ineffective for rare or novel diseases where patient samples are inherently scarce. RESULTS: Here, we present ProMeta, a meta-learning framework designed to enable robust disease modeling under extreme data restrictions. By integrating knowledge-guided pathway encoding with bi-level meta-optimization, ProMeta projects unstructured proteomic profiles into biologically interpretable functional tokens. This architecture allows the model to learn a global initialization containing transferable biological priors from biobank-scale data, facilitating rapid adaptation to novel tasks. Through comprehensive benchmark experiments, ProMeta consistently outperformed transfer learning and traditional machine learning baselines in both disease diagnosis and prediction tasks. In the most challenging 4-shot scenarios (utilizing only 2 cases and 2 controls), the model achieved robust generalization with an average AUROC of ∼0.69, representing a 24.6% relative improvement over the best-performing baseline methods. Mechanistic investigation revealed that ProMeta disentangles cases from controls in the latent space prior to task-specific adaptation, confirming the acquisition of universal biological rules rather than rote memorization. Furthermore, gradient-based interpretation identified disease-specific protein biomarkers and functional pathways consistent with known pathophysiology. Collectively, ProMeta overcomes the data-scarcity bottleneck in precision medicine, providing a scalable, interpretable framework for characterizing the full spectrum of human diseases, particularly for rare conditions lacking extensive clinical cohorts. AVAILABILITY AND IMPLEMENTATION: The source code of ProMeta is available at GitHub (https://github.com/lihan97/ProMeta).

Proteomics↗

PubMed2026

Inclusion of Multi-Omic Biomarkers Improves Prediction Accuracy of Response, Relapse, and Overall Survival in Acute Myeloid Leukemia Patients Receiving High-Intensity Induction Chemotherapy.

BACKGROUND: Despite advancements in genetic markers for acute myeloid leukemia (AML) risk stratification, outcome prediction remains challenging due to disease heterogeneity and dynamic genetic changes, highlighting the need for reliable biomarkers to improve AML treatment strategies and patient outcomes. To refine outcome predictions, we investigated the use of microbial-derived biomarkers to predict composite complete remission (CRc), relapse, and survival for patients on high- and low-intensity regimens, and to integrate those variables into the widely clinically utilized European Leukemia Network (ELN-2022) genetic risk classification model for high-intensity-treated patients. METHODS: We first developed machine learning models that integrate baseline fecal metabolomics, 16S rRNA-based stool microbiome features, and clinical metadata (sex, antibiotic administration, AML somatic mutations, and cytogenetics) from two cohorts of AML patients (n = 83) undergoing remission induction chemotherapy. Univariate tests and sparse canonical correlation analysis were employed for variable selection and to explore fecal metabolite-microbe relationships. A robust machine learning approach using XGBoost was employed, with 100 stratified data splits (80% training, 20% testing) and coarse-to-fine hyperparameter optimization. Variable importance was aggregated across all models to select key predictors. RESULTS: For high-intensity-treated patients, XGBoost models achieved aggregated AUROC scores of 0.719, 0.729, and 0.65 for CRc, relapse, and overall survival, respectively. For low-intensity-treated patients, these models achieved aggregate AUROC scores of 0.945, 0.724, and 0.768 for these same outcomes, respectively. Integrating the biomarkers identified in the high-intensity machine-learning models with the current ELN-2022 AML risk stratification system effectively stratified patients into risk categories, which obtained higher concordance indices and likelihood ratios, demonstrating improved prognostic accuracy for each outcome compared to ELN-2022 alone. CONCLUSIONS: The inclusion of microbial-derived biomarkers serves as a robust prognostic tool to improve outcome prediction in AML patients, highlighting the potential of its integration into AML risk assessment and paving the way for personalized treatment strategies and improved patient outcomes.

Humans↗

PubMed2026

Discovery and validation of a multi-protein panel for predicting non-fatal major adverse cardiovascular events in diabetic kidney disease.

OBJECTIVE: To identify plasma protein biomarkers associated with incident non-fatal major adverse cardiovascular events (MACE) in diabetic kidney disease (DKD) patients. RESEARCH DESIGN AND METHODS: We analyzed 317 DKD patients from the UK Biobank. Plasma proteomics and clinical data (demographics, metabolism, renal function) were integrated. In an exploratory discovery phase, three sequential Cox regression models (crude, socio-demographic-adjusted, socio-demographic-metabolic adjusted) screened non-fatal MACE-associated proteins. To prevent information leakage, the cohort was then randomly split into training (70%) and testing (30%) sets; machine-learning feature selection, hyperparameter optimization, and final model development were performed exclusively within the training set. The associated proteins were input into the four-step machine-learning pipeline (LASSO-Cox, random survival forest, Boruta, XGBoost-Cox). Predictive performance was validated using Kaplan-Meier survival analyses, longitudinal trajectory modeling, and ROC benchmarking. An interactive web application was deployed for clinical implementation. RESULTS: Of 1,463 plasma proteins, 561 were associated with non-fatal MACE across Cox models, with 14 overlapping proteins. Nine core proteins (ANG, IL1R1, CXCL14, ESAM, PTGDS, HAVCR1, FGFR2, IGSF8, CCL3) were validated: ANG showed the strongest non-fatal MACE association (HR = 3.88, 95%CI 2.33-6.48, p<0.001), and all high-expression groups had elevated non-fatal MACE risk. GO/KEGG enrichment highlighted inflammatory-immune pathways like positive regulation of MAPK cascade, Cytokine-cytokine receptor interaction and PI3K-Akt signaling pathway as key mechanisms. The model integrating proteins, demographic factors, and clinical variables achieved the highest predictive performance across non-fatal MACE (AUC = 0.768), myocardial infarction (MI) (0.808), and stroke (0.816) outcomes, with superior stability in cross-validation. CoxBoost + Elastic Net framework was selected as the optimal framework via benchmarking of 101 algorithms. The model demonstrated favorable calibration in high-risk patients and yielded positive net clinical benefit across decision thresholds of 5% to 45%. The web tool (https://jiangli2941.github.io/MACE-prediction-v2/) enables input of 28 variables, outputs non-fatal MACE risk status, risk probability, and highlights abnormal indicators. CONCLUSION: Plasma proteomics combined with machine learning identifies robust non-fatal MACE predictors in DKD.

Humans↗

PubMed2026

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 688 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.

DNA methylation↗

PubMed2025

Disease candidate genes prediction using positive labeled and unlabeled instances.

Identifying disease genes and understanding their performance is critical in producing drugs for genetic diseases. Nowadays, laboratory approaches are not only used for disease gene identification but also using computational approaches like machine learning are becoming considerable for this purpose. In machine learning methods, researchers can only use two data types (disease genes and unknown genes) to predict disease candidate genes. Notably, there is no source for the negative data set. The proposed method is a two-step process: The first step is the extraction of reliable negative genes from a set of unlabeled genes by one-class learning and a filter based on distance indicators from known disease genes; this step is performed separately for each disease. The second step is the learning of a binary model using causing genes of each disease as a positive learning set and the reliable negative genes extracted from that disease. Each gene in the unlabeled gene's production and ranking step is assigned a normalized score using two filters and a learned model. Consequently, disease genes are predicted and ranked. The proposed method evaluation of various six diseases and Cancer class indicates better results than other studies.

Humans↗

PubMed2026

Meta-PseU: A meta-classifier for robust prediction of RNA pseudouridine modification sites from long sequences.

BACKGROUND AND OBJECTIVES: Pseudouridine (Ψ) represents one of the most abundant and conserved RNA modifications. Ψ provides an additional hydrogen-bond donor that enhances RNA structural stability and modulates translation. It participates in diverse biological processes, including RNA-protein interactions, splicing, translational control, and stress responses. Aberrant pseudouridylation is implicated in cancer, neurodegenerative disorders, and autoimmune diseases. Despite its biological importance, experimental identification of Ψ sites remains time-consuming and costly, limiting the feasibility of transcriptome-wide profiling. Computational approaches have therefore become essential complements to experimental techniques. However, state-of-the-art machine-learning and deep-learning predictors often suffer from limited generalizability due to small training datasets. To overcome these issues, we aim at constructing new long-sequence datasets and developing a novel Ψ site predictor. METHODS: New long-sequence datasets were constructed as benchmarks for RNA Ψ-site prediction. The Ψ modification sites in RMBase 3.0 were mapped to the reference genomes across three species of human, mouse, and yeast, and the RNA sequences with a length of 201 were generated by extending the upstream and downstream from the mapped, central sites. To eliminate sequence redundancy, the sequences were clustered using CD-HIT with a 70% sequence identity threshold. We developed Meta-PseU, a logistic regression-based meta-classifier that considered 118 machine learning and deep learning classifiers. The datasets and programs are freely accessible at https://github.com/kuratahiroyuki/MetaPseU. RESULTS: By optimizing model configuration, we proposed the Meta-PseU model stacking 32 machine learning and deep learning classifiers out of 118 classifiers. Meta-PseU substantially improved model generalizability, overcoming a key limitation of existing approaches. It greatly outperformed state-of-the-art predictors and achieved increasing accuracy with increasing sequence length. CONCLUSIONS: Long-sequence datasets were newly constructed as benchmarks for RNA Ψ-site prediction. Meta-PseU offers a new framework for robust Ψ-site identification by using long sequences.

Pseudouridine↗

PubMed2025

Structure-based discovery of inhibitors of Mac1 domain of nonstructural protein-3 of SARS-CoV-2 by machine learning-augmented screening of chemical space.

Significant efforts have been recently dedicated to the discovery of small molecule inhibitors against the Macrodomain 1 (Mac1) of nonstructural protein 3 (NSP3) as potential antivirals for SARS-CoV-2. Thus, Mac1 has also been selected as the target for the Critical Assessment of Hit-finding Experiments (CACHE) challenge #3. As contestants in that challenge, we developed a computational strategy that ranked on the top among all 23 participants in the competition and resulted in the discovery of a novel chemical series of non-charged Mac1 inhibitors. Those have been identified through the combination of machine learning-accelerated virtual screening of Enamine REAL Diversity Subset of approximately 25 million compounds and consequent hit expansion into the entire Enamine REAL Space library. In particular, the initially identified hit compound CACHE3-HI_1706_56 (KD = 20 μM) was explored by probing 17 close analogues from a library of 44 billion molecules from the Enamine REAL. All those analogues effectively displaced the Mac1-binding ADP-ribose peptide, and 12 were confirmed to engage with Mac1 by the Surface Plasmon Resonance experiments, revealing a new chemical series of compounds for hit-to-lead optimization. The structure of the CACHE3-HI_1706_56-Mac1 complex was further determined at high resolution with crystallography, confirming initial computational predictions. Our results illustrate the effectiveness of ML-accelerated docking to rapidly identify novel chemical series and provide a strong foundation for the development of SARS-CoV-2 NSP3 Mac1 inhibitors.

CACHE challenge↗

PubMed2026

Machine learning-assisted Mn-N-C nanozyme colorimetric sensor array for trace-level detection of biogenic amines in meat.

Accurate detection of biogenic amines (BAs) in meat remains challenging due to their high structural similarity and co-occurrence. Herein, an Mn-N-C nanozyme was synthesized via a metal-organic framework confined pyrolysis strategy, possessing excellent oxidase (OXD)- and peroxidase (POD)-like activities. The dual enzyme-like activity showed Km values of 0.1584 mM (OXD) and 0.1498 mM (POD), respectively, in detection system. Leveraging these properties, a colorimetric sensor array was constructed, enabling the detection of four representative BAs within a concentration range of 2-10 ppm with 100% classification accuracy. In addition, a concentration independent recognition model based on an artificial neural network was developed to address signal nonlinearity interference in meat. The integrated system achieved accurate trace-level identification of BAs in perishable fish, pork, and chicken, demonstrating its applicability for early-stage BAs monitoring and quality deterioration warning during storage and transportation.

Biogenic Amines↗

PubMed2026

Machine Learning-Driven Prediction of Coronary Artery Disease Risk Based on UK Biobank Plasma Proteomics.

BACKGROUND: Coronary artery disease (CAD) is a leading global cause of mortality, yet the predictive accuracy of conventional risk models is limited. Here, we integrate conventional risk factors, polygenic risk scores, and large-scale proteomics to develop a unified model for enhanced CAD risk prediction. METHODS: Using data from UK Biobank, participants with plasma proteomics and genetic risk data were included after excluding prevalent CAD. Participants from England were split into training (n=32 330) and internal validation (n=13 857) sets, and Scotland/Wales participants formed an external validation set (n=5775). Incident CAD was ascertained from linked health records. A 202-protein proteomic risk score was derived by least absolute shrinkage and selection operator Cox regression, and CatBoost models were trained using conventional risk factors alone and with incremental addition of polygenic risk scores and protein proteomic risk scores; Shapley Additive Explanations-guided forward selection identified a compact protein panel. RESULTS: Across cohorts, the median age was 58 years and ∼45% were men. Protein proteomic risk score was dose-dependently associated with CAD risk. Compared with conventional risk factors alone, integrating polygenic risk scores and protein proteomic risk scores improved discrimination, with the area under the curve increasing from 0.750 (95% CI, 0.732-0.767) to 0.789 (95% CI, 0.772-0.805) in internal validation and from 0.717 (95% CI, 0.683-0.750) to 0.762 (95% CI, 0.732-0.791) in external validation. A 9-protein panel (GDF15 [growth differentiation factor 15], MMP12 [matrix metalloproteinase 12], NPPB [natriuretic peptide B], PGF [placental growth factor], REN [renin], ADGRG2 [adhesion G-protein coupled receptor], ACE2 [angiotensin-converting enzyme 2], CDCP1 [CUB domain-containing protein 1], CXCL17 [C-X-C motif chemokine ligand 17)]) captured most proteomic predictive information. CONCLUSIONS: Our findings demonstrate that integrating conventional risk factors, polygenic risk scores, and proteomic data improves CAD risk prediction. This study highlights the utility of proteomics in precision cardiovascular medicine and simplified risk stratification tools.

Humans↗

PubMed2025

Data-driven consideration of genetic disorders for global genomic newborn screening programs.

PURPOSE: Over 30 international studies are exploring newborn sequencing (NBSeq) to expand the range of genetic disorders included in newborn screening. Substantial variability in gene selection across programs exists, highlighting the need for a systematic approach to prioritize genes. METHODS: We assembled a data set comprising 25 characteristics about each of the 4390 genes included in 27 NBSeq programs. We used regression analysis to identify several predictors of inclusion and developed a machine learning model to rank genes for public health consideration. RESULTS: Among 27 NBSeq programs, the number of genes analyzed ranged from 134 to 4299, with only 74 (1.7%) genes included by over 80% of programs. The most significant associations with gene inclusion across programs were presence on the US Recommended Uniform Screening Panel (inclusion increase of 74.7%, CI: 71.0%-78.4%), robust evidence on the natural history (29.5%, CI: 24.6%-34.4%), and treatment efficacy (17.0%, CI: 12.3%-21.7%) of the associated genetic disease. A boosted trees machine learning model using 13 predictors achieved high accuracy in predicting gene inclusion across programs (area under the curve = 0.915, R2 = 84%). CONCLUSION: The machine learning model developed here provides a ranked list of genes that can adapt to emerging evidence and regional needs, enabling more consistent and informed gene selection in NBSeq initiatives.

Humans↗

PubMed2025

Identification and validation of the important role of KIF11 in the development and progression of endometrial cancer.

BACKGROUND: Human kinesin family member 11 (KIF11) plays a vital role in regulating the cell cycle and is implicated in the tumorigenesis and progression of various cancers, but its role in endometrial cancer (EC) is still unclear. Our current research explored the prognostic value, biological function and targeting strategy of KIF11 in EC through approaches including bioinformatics, machine learning and experimental studies. METHODS: The GSE17025 dataset from the GEO database was analyzed via the limma package to identify differentially expressed genes (DEGs) in EC. Functional enrichment analysis of the DEGs was conducted using Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses. DEGs were further screened for hub genes through protein-protein interaction (PPI) network analysis and machine learning. The role of the hub gene KIF11 in EC was analyzed using clinical data from the TCGA database. The expression of KIF11 in EC was subsequently validated in clinical samples. In vitro experiments were utilized to evaluate the effects of KIF11 on biological functions such as proliferation, migration, apoptosis, and the cell cycle in endometrial cancer cells. RESULTS: A total of 877 DEGs, which are widely involved in important biological processes such as cell division, tubulin binding, and the cell cycle, were identified. Through PPI network analysis and machine learning, KIF11 was selected as the hub gene for subsequent analysis and experimental validation. An analysis of TCGA data revealed that KIF11 is highly expressed in EC and is associated with tumor grade, stage, and a low survival rate. The overexpression of KIF11 in tumor tissues was further confirmed in EC patient samples. KIF11 knockdown had inhibitory effects on cell proliferation, migration and invasion. Flow cytometry analysis revealed that KIF11 knockdown induced G2/M phase arrest and promoted apoptosis in EC cells. CONCLUSION: Our study demonstrated that KIF11 was upregulated in EC and was strongly associated with a poor prognosis. Notably, we found that reduced KIF11 expression inhibited EC cell proliferation, migration and invasion. KIF11 knockdown caused more EC cells to arrest in the G2/M phase and undergo apoptosis. The findings of our study emphasized that KIF11 may be a promising prognostic biomarker and therapeutic target for EC patients.

Humans↗

PubMed2025

Predicting host tropism in influenza a viruses: insights from multi-segment nucleotide signatures.

BACKGROUND: Influenza A virus (IAV) poses a significant public health threat due to its cross-species transmission and complex host adaptation mechanisms. This study integrated whole-genome data from avian, human, swine, and bovine IAV strains, using machine learning to predict viral host tropism based on nucleotide site features and to identify key sites driving host adaptation along with their synergistic effects. METHODS: A total of 64,000 IAV sequences from avian, human, swine, and bovine hosts were analyzed to build host-prediction models. A four-class classification framework (avian, human, swine, bovine) was constructed using nucleotide site features from all eight genomic segments (PB2, PB1, PA, HA, NP, NA, MP, NS). Eight machine learning algorithms (logistic regression, decision tree, random forest, SVM, KNN, gradient boosting, XGBoost, LightGBM) were benchmarked via 10-fold stratified cross-validation. Model performance was evaluated using accuracy, precision, recall, F1-score, AUPRC, and AUC. SHAP (SHapley Additive exPlanations) analysis prioritized critical nucleotide sites, while bivariate association tests identified synergistic/antagonistic interactions between sites. Nucleotide composition profiles were compared across host groups using hierarchical clustering and heatmap visualization. RESULTS: The XGBoost algorithm demonstrated the best and most stable performance, achieving an AUC value of over 0.95 in distinguishing human-derived sequences from non-human ones. SHAP analysis identified the top 20 critical nucleotide sites for each gene segment, such as sites 46 and 698 in the NS segment. Nucleotide composition analysis revealed high similarity between human and swine sequences in the HA and PB2 segments, and between avian and bovine sequences. The HA segment was particularly challenging in differentiating human from swine strains. Bivariate site association analysis uncovered significant synergistic or antagonistic effects between key sites within gene segments, forming complex networks. For instance, in the NS segment, a positive prediction contribution was observed when sites 371, 698, and 419 were all G. CONCLUSIONS: This study advances our mechanistic understanding of IAV host adaptation, identifies molecular determinants for zoonotic risk stratification, and establishes a scalable machine learning framework for predicting viral host tropism through nucleotide signature analysis, thereby enhancing surveillance strategies and informing preventive measures against emerging viral threats.

Influenza A virus↗

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

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↗

PubMed1993

Prediction of primate splice junction gene sequences with a cooperative knowledge acquisition system.

We propose a cooperative conceptual modelling environment in which two agents interact: the machine and the human expert. The former is able to extract knowledge from data using a symbolic-numeric machine learning system, and the latter is able to control the learning process by accepting and validating the machine results, or by criticizing those results or the explanation that the system produces on them. The improvement of the conceptual modelling relies on the cooperation between the two agents. Results obtained with our method on prediction of primate splice junctions sites in genetic sequences are far better than those reported in the literature with other symbolic machine learning systems, and are as better as those obtained with some artificial neural networks methods reported at present. But in opposite to neural networks which lack of argumentation, our system provides the user a plausible explanation of its prediction.

Algorithms↗

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↗

PubMed2026

Machine learning-driven spleen imaging and genomics uncover a splenic connection to coronary artery disease.

Despite advances in managing traditional risk factors, coronary artery disease (CAD) remains the leading cause of mortality. Circulating hematopoietic cells influence risk for CAD separately from traditional risk factors, but the role of a key regulating organ, the spleen, is unknown. The understudied spleen is a representation of the hematopoietic system optimally suited for unbiased radiologic investigations toward mechanistic insights. Here, we leveraged deep learning to extract 107 splenic radiomic features from abdominal magnetic resonance imaging (MRI) scans of 42,059 UK Biobank participants and of 2745 Mass General Brigham Biobank (MGBB) participants. Of these, 10 features from UK Biobank were associated with CAD. Genome-wide association analysis of CAD-associated features identified 219 loci, including 9p21. Variants at 9p21, the strongest yet mechanistically elusive CAD locus, were associated with splenic features such as run-length nonuniformity, reflecting heterogeneity of continuous texture regions. Research MRI findings were consistent internally, but external clinical validation highlighted challenges in translating analyses of abdominal MRI scans to routine clinical practice because of variability in imaging protocols and greater clinical heterogeneity among patients. Our study, combining deep learning with genomics, presents a framework to uncover potential splenic involvement in CAD and emphasizes translational gaps between research and clinical radiomics.

Humans↗

PubMed2026

Predicting cellular responses to perturbation across diverse contexts with State.

While machine learning models offer potential for predicting transcriptomic effects of perturbation, they currently struggle to generalize across cellular contexts. Here, we introduce State, a machine learning model that predicts perturbation effects while accounting for cellular heterogeneity within and across experiments. State is trained using single-cell gene expression data to predict perturbation effects across sets of cells. State improved discrimination of effects on large datasets by more than 30% and identified differentially expressed genes across genetic, signaling, and chemical perturbations with significantly improved accuracy compared with baselines. Its cell embeddings trained on observational data from 167 million cells enable the identification of strong perturbations in cellular contexts where no perturbations were observed during training. We further introduce Cell-Eval, a comprehensive evaluation framework that can be used to evaluate future models. Overall, the performance and flexibility of State set the stage for scaling the development of AI models of cell state.

Machine Learning↗