Computational biology and chemistry · 2026
AI-driven transcriptomics for neurodegenerative disease research: A systematic review.
СтатьяОбзорНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Tripathi Anushree, Yadav Anoop Kumar, Kumar Nilesh
Аннотация
BACKGROUND: Neurodegenerative diseases (NDDs), including Alzheimer's disease (AD), Parkinson's disease (PD), and Amyotrophic Lateral Sclerosis (ALS), present a growing global health challenge, with traditional diagnostic and therapeutic approaches facing significant limitations in early detection and mechanistic resolution.
OBJECTIVES: Innovations in transcriptomic technologies-spanning bulk RNA sequencing, single-cell RNA-seq (scRNA-seq), and spatially resolved transcriptomics-provide unprecedented insights into cellular heterogeneity and microenvironmental dysregulation. This systematic review synthesizes recent applications of artificial intelligence (AI), machine learning (ML), and deep learning (DL) across transcriptomic tiers to benchmark computational models for biomarker identification, disease diagnosis, patient subtyping, and drug repurposing.
METHODS & ELIGIBILITY: Following PRISMA 2020 guidelines for qualitative systematic reviews, literature published between 2016 and 2025 was systematically searched across PubMed/MEDLINE, IEEE Xplore, Google Scholar, and bioRxiv/medRxiv. Of 1058 records identified, 19 core empirical studies met all eligibility criteria and were synthesized in depth alongside contextual reference benchmarks.
SYNTHESIS & KEY FINDINGS: Traditional ML algorithms (SVM, RF, Elastic Net) demonstrate high diagnostic accuracy (AUC 0.72-0.98) in blood-based biomarker selection, while deep architectures (autoencoders, LSTMs, and Graph Neural Networks) excel at modeling continuous disease trajectories and resolving spatial transcriptomic niches (e.g., STAGATE, SpaGCN, cell2location). Emerging spatial deep learning models successfully localize disease-associated microglia (DAM) around amyloid plaques and trace axonal degeneration pathways.
LIMITATIONS & CONCLUSIONS: Key translational hurdles include data scarcity in rare NDD subtypes, lack of cross-platform spatial benchmarks, and the black-box nature of deep neural networks. Integrating explainable AI (XAI) and prospective multi-modal cohorts is essential for translating computational transcriptomics into clinical diagnostics and targeted therapeutics.
OBJECTIVES: Innovations in transcriptomic technologies-spanning bulk RNA sequencing, single-cell RNA-seq (scRNA-seq), and spatially resolved transcriptomics-provide unprecedented insights into cellular heterogeneity and microenvironmental dysregulation. This systematic review synthesizes recent applications of artificial intelligence (AI), machine learning (ML), and deep learning (DL) across transcriptomic tiers to benchmark computational models for biomarker identification, disease diagnosis, patient subtyping, and drug repurposing.
METHODS & ELIGIBILITY: Following PRISMA 2020 guidelines for qualitative systematic reviews, literature published between 2016 and 2025 was systematically searched across PubMed/MEDLINE, IEEE Xplore, Google Scholar, and bioRxiv/medRxiv. Of 1058 records identified, 19 core empirical studies met all eligibility criteria and were synthesized in depth alongside contextual reference benchmarks.
SYNTHESIS & KEY FINDINGS: Traditional ML algorithms (SVM, RF, Elastic Net) demonstrate high diagnostic accuracy (AUC 0.72-0.98) in blood-based biomarker selection, while deep architectures (autoencoders, LSTMs, and Graph Neural Networks) excel at modeling continuous disease trajectories and resolving spatial transcriptomic niches (e.g., STAGATE, SpaGCN, cell2location). Emerging spatial deep learning models successfully localize disease-associated microglia (DAM) around amyloid plaques and trace axonal degeneration pathways.
LIMITATIONS & CONCLUSIONS: Key translational hurdles include data scarcity in rare NDD subtypes, lack of cross-platform spatial benchmarks, and the black-box nature of deep neural networks. Integrating explainable AI (XAI) and prospective multi-modal cohorts is essential for translating computational transcriptomics into clinical diagnostics and targeted therapeutics.