Long non-coding RNAs in human peripheral blood as biomarkers for age prediction

Chronological age is a critical biological trait with substantial relevance in forensic investigations. Growing evidence highlights the role of long non-coding RNAs (lncRNAs) in aging, but the application in age prediction remains largely unexplored. To evaluate the potential of lncRNAs for age prediction, we analyzed RNA-seq data from peripheral blood samples of Han Chinese individuals. Differential expression analysis identified a series of differentially expressed lncRNAs (DElncRNAs) among multiple age group comparisons, with the most significant expression changes observed between the youngest (18-27 years) and older (68-77 years) age groups. GO and KEGG enrichment analysis of target genes indicated that these lncRNAs were involved in mitochondrial function, telomere maintenance, and pathways related to neurodegenerative disease. Additionally, Spearman's correlation analysis further identified 20 lncRNAs significantly associated with age as candidate markers, which were subsequently used to construct age prediction models using six machine learning algorithms. The XGBoost model achieved the best performance, with a mean absolute error (MAE) of 8.04 years (R² = 0.66) in the test cohort. External validation was conducted using two independent datasets (GSE124326 and GSE177044). The XGBoost and LightGBM models consistently outperformed the other models, with MAEs ranging from 7.99 to 8.21 years. These findings demonstrated the feasibility of lncRNA-based age prediction, providing novel biomarkers and methodological insights for forensic age estimation.