Mia Yang Ang, Li Chen, Lanni Song, Leonard Lipovich, Siew Woh Choo
This review synthesizes literature on GWAS interpretation, epigenomic annotation, single-cell and spatial transcriptomics, multi-omics integration, artificial intelligence and machine learning, experimental validation, RNA therapeutic modality selection, delivery assessment, safety evaluation, and biomarker-informed precision medicine to present a framework for AI-guided lncRNA target discovery. Genetic and multi-omics data can nominate disease-relevant lncRNAs, but no single evidence layer is sufficient to establish causality, mechanism, druggability, or clinical utility. AI can integrate heterogeneous biomedical datasets, rank candidate lncRNAs, detect regulatory patterns, and prioritize transcripts for validation. LncRNAs represent a potential therapeutic target space for precision medicine, but candidate targets require disease-context expression validation, functional perturbation, mechanistic assessment, appropriate model systems, therapeutic modulation, delivery-feasibility assessment, safety evaluation, and patient-selection strategies.
Background/Objectives: Precision medicine requires translation of genetic and molecular variation into clinically actionable therapeutic targets. However, many disease-associated signals identified by genome-wide association studies (GWAS) reside in non-coding regulatory regions, making biological interpretation and therapeutic prioritization difficult. Long non-coding RNAs (lncRNAs) are regulatory molecules with growing relevance to disease mechanisms, biomarker discovery, patient stratification, and RNA-based therapeutics. This review presents a translational framework for AI-guided lncRNA target discovery, linking non-coding genetic signals to experimental validation and clinical implementation. Methods: This narrative review synthesizes literature on GWAS interpretation, quantitative trait loci analysis, epigenomic annotation, single-cell and spatial transcriptomics, multi-omics integration, artificial intelligence and machine learning, experimental validation, RNA therapeutic modality selection, delivery assessment, safety evaluation, and biomarker-informed precision medicine. Results: Genetic and multi-omics data can nominate disease-relevant lncRNAs, but no single evidence layer is sufficient to establish causality, mechanism, druggability, or clinical utility. AI can integrate heterogeneous biomedical datasets, rank candidate lncRNAs, detect regulatory patterns, and prioritize transcripts for validation. However, computational prediction should be interpreted as decision support rather than proof of therapeutic relevance. Candidate targets require disease-context expression validation, functional perturbation, mechanistic assessment, appropriate model systems, therapeutic modulation, delivery-feasibility assessment, safety evaluation, and patient-selection strategies. Conclusions: LncRNAs represent a promising but challenging therapeutic target class. A responsible translational pipeline should connect non-coding genetic evidence and multi-omics support with AI-guided prioritization, experimental validation, RNA therapeutic strategy selection, delivery assessment, safety evaluation, and clinical implementation. The framework defines qualification criteria and decision gates to reduce premature target claims during translational development.