logo logo
A versatile model for single-cell data analysis. Nature computational science 10.1038/s43588-021-00103-1
Single-Cell Analysis Using Machine Learning Techniques and Its Application to Medical Research. Asada Ken,Takasawa Ken,Machino Hidenori,Takahashi Satoshi,Shinkai Norio,Bolatkan Amina,Kobayashi Kazuma,Komatsu Masaaki,Kaneko Syuzo,Okamoto Koji,Hamamoto Ryuji Biomedicines In recent years, the diversity of cancer cells in tumor tissues as a result of intratumor heterogeneity has attracted attention. In particular, the development of single-cell analysis technology has made a significant contribution to the field; technologies that are centered on single-cell RNA sequencing (scRNA-seq) have been reported to analyze cancer constituent cells, identify cell groups responsible for therapeutic resistance, and analyze gene signatures of resistant cell groups. However, although single-cell analysis is a powerful tool, various issues have been reported, including batch effects and transcriptional noise due to gene expression variation and mRNA degradation. To overcome these issues, machine learning techniques are currently being introduced for single-cell analysis, and promising results are being reported. In addition, machine learning has also been used in various ways for single-cell analysis, such as single-cell assay of transposase accessible chromatin sequencing (ATAC-seq), chromatin immunoprecipitation sequencing (ChIP-seq) analysis, and multi-omics analysis; thus, it contributes to a deeper understanding of the characteristics of human diseases, especially cancer, and supports clinical applications. In this review, we present a comprehensive introduction to the implementation of machine learning techniques in medical research for single-cell analysis, and discuss their usefulness and future potential. 10.3390/biomedicines9111513
Comprehensive multi-omics integration uncovers mitochondrial gene signatures for prognosis and personalized therapy in lung adenocarcinoma. Journal of translational medicine The therapeutic efficacy of lung adenocarcinoma (LUAD), the most prevalent histological subtype of primary lung cancer, remains inadequate, with accurate prognostic assessment posing significant challenges. This study sought to elucidate the prognostic significance of mitochondrial-related genes in LUAD through an integrative multi-omics approach, aimed at developing personalized therapeutic strategies. Utilizing transcriptomic and single-cell RNA sequencing (scRNA-seq) data, alongside clinical information from publicly available databases, we first applied dimensionality reduction and clustering techniques to the LUAD single-cell dataset, focusing on the subclassification of fibroblasts, epithelial cells, and T cells. Mitochondrial-related prognostic genes were subsequently identified using TCGA-LUAD data, and LUAD cases were stratified into distinct molecular subtypes through consensus clustering, allowing for the exploration of gene expression profiles and clinical feature distributions across subtypes. By leveraging an ensemble of machine learning algorithms, we developed an Artificial Intelligence-Derived Prognostic Signature (AIDPS) model based on mitochondrial-related genes and validated its prognostic accuracy across multiple independent datasets. The AIDPS model demonstrated robust predictive power for LUAD patient outcomes, revealing significant differences in responses to immunotherapy and chemotherapy, as well as survival outcomes between risk groups. Furthermore, we conducted comprehensive analyses of tumor mutation burden (TMB), immune microenvironment characteristics, and genome-wide association study (GWAS) data, providing additional insights into the mechanistic roles of mitochondrial-related genes in LUAD pathogenesis. This study not only offers a novel approach to improving prognostic assessments in LUAD but also establishes a strong foundation for the development of personalized therapeutic interventions. 10.1186/s12967-024-05754-y
Early detection of Dementia in Type 2 Diabetes population: Predictive analytics using Machine learning approach. Journal of medical Internet research BACKGROUND:The possible association between diabetes mellitus and dementia has raised concerns, given the observed coincidental occurrences. OBJECTIVE:This study aims to develop a personalized predictive model, utilizing artificial intelligence, to assess the 5-year and 10-year dementia risk among patients with Type 2 Diabetes Mellitus (T2DM) who are prescribed antidiabetic medications. METHODS:This retrospective multicenter study used data from Taipei Medical University Clinical Research Database, which comprises electronic medical records from three hospitals in Taiwan. This study applied eight machine learning algorithms to develop prediction models, including logistic regression (LR), linear discriminant analysis (LDA), gradient boosting machine (GBM), lightGBM (LBGM), AdaBoost, random forest, extreme gradient boosting (XGBoost), and artificial neural network (ANN). These models incorporated a range of variables, encompassing patient characteristics, comorbidities, medication usage, laboratory results, and examination data. RESULTS:This study involved a cohort of 43,068 patients diagnosed with T2DM, which accounted for a total of 1,937,692 visits. For model development and validation, 1,300,829 visits were utilized, while an additional 636,863 visits were reserved for external testing. The area under the curve (AUC) of the prediction models range from 0.67 for the logistic regression to 0.98 for the artificial neural networks. Based on the external test results, the model built using the ANN algorithm has the best AUC: 0.97 (5-year follow-up period) and 0.98 (10-year follow-up period). Based on the best model (ANN), age, gender, triglyceride, HbA1c, anti-diabetic agents, stroke history, and other long-term medications were the most important predictors. CONCLUSIONS:We have successfully developed a novel computer-aided dementia risk prediction model that can facilitate the clinical diagnosis and management of patients prescribed with antidiabetic medications. However, further investigation is required to assess the model's feasibility and external validity. CLINICALTRIAL: 10.2196/52107
Development of a prognostic model related to homologous recombination deficiency in glioma based on multiple machine learning. Frontiers in immunology Background:Despite advances in neuro-oncology, treatments of glioma and tools for predicting the outcome of patients remain limited. The objective of this research is to construct a prognostic model for glioma using the Homologous Recombination Deficiency (HRD) score and validate its predictive capability for glioma. Methods:We consolidated glioma datasets from TCGA, various cancer types for pan-cancer HRD analysis, and two additional glioma RNAseq datasets from GEO and CGGA databases. HRD scores, mutation data, and other genomic indices were calculated. Using machine learning algorithms, we identified signature genes and constructed an HRD-related prognostic risk model. The model's performance was validated across multiple cohorts. We also assessed immune infiltration and conducted molecular docking to identify potential therapeutic agents. Results:Our analysis established a correlation between higher HRD scores and genomic instability in gliomas. The model, based on machine learning algorithms, identified seven key genes, significantly predicting patient prognosis. Moreover, the HRD score prognostic model surpassed other models in terms of prediction efficacy across different cancers. Differential immune cell infiltration patterns were observed between HRD risk groups, with potential implications for immunotherapy. Molecular docking highlighted several compounds, notably Panobinostat, as promising for high-risk patients. Conclusions:The prognostic model based on the HRD score threshold and associated genes in glioma offers new insights into the genomic and immunological landscapes, potentially guiding therapeutic strategies. The differential immune profiles associated with HRD-risk groups could inform immunotherapeutic interventions, with our findings paving the way for personalized medicine in glioma treatment. 10.3389/fimmu.2024.1452097
Discovery of High-Risk Clinical Factors That Accelerate Brain Aging in Adults: A Population-Based Machine Learning Study. Research (Washington, D.C.) Brain age prediction using neuroimaging data and machine learning algorithms holds significant promise for gaining insights into the development of neurodegenerative diseases. The estimation of brain age may be influenced not only by the imaging modality but also by multidomain clinical factors. However, the degree to which various clinical factors in individuals are associated with brain structure, as well as the comprehensive relationship between these factors and brain aging, is not yet clear. In this study, multimodal brain magnetic resonance imaging data and longitudinal clinical information were collected from 964 participants in a population-based cohort with 16 years of follow-up in northern China. We developed a machine learning-based algorithm to predict multimodal brain age and compared the estimated brain age gap (BAG) differences among the 5 groups characterized by varying exposures to these high-risk clinical factors. We then estimated modality-specific brain age in the hypertension group based on hypertension-related regional imaging metrics. The results revealed a significantly larger BAG estimated from multimodal neuroimaging in subjects with 4 or 5 risk factors compared to other groups, suggesting an acceleration of brain aging under cumulative exposure to multiple risk factors. The estimated T1-based BAG exhibited a significantly higher level in the hypertensive subjects compared to the normotensive individuals. Our study provides valuable insights into a range of health factors across lifestyle, metabolism, and social context that are reflective of brain aging and also contributes to the advancement of interventions and public health initiatives targeted at the general population aimed at promoting brain health. 10.34133/research.0500