Scholarlink

<p>With advances in transcriptomics, molecular markers identified at the RNA level are being widely explored in age prediction. RNA-seq reads allow for reliable quantification of transcript-level expression by next-generation sequencing (NGS) platforms. Given that post-transcriptional regulation plays a critical role in aging, transcript isoforms and their expression may serve as novel biomarkers for age prediction. In this study, a total of 171,838 transcripts were identified from the transcriptome expression profiles of blood samples from 127 Han Chinese individuals. Differential expression analysis identified 590 transcripts with age-related expression changes across different age groups, and 16 age-related transcripts were identified by Spearman correlation analysis. Four age prediction models were constructed using these transcripts, and their performance was evaluated with an independent test set comprising 150 samples. Among these models, the Extreme Gradient Boosting (XGBoost) model exhibited the optimal performance, with a mean absolute error (MAE) of 7.609 years. Furthermore, explainable Shapley Additive exPlanations (SHAP) analysis was performed on this optimal XGBoost model, which identified four transcripts with negligible contributions to model performance. Ultimately, a final refined age prediction model constructed from the remaining 12 transcripts achieved an MAE of 7.714 years. Overall, this study investigated the potential of transcript expression levels in age prediction, providing new biomarkers and methodological insights for forensic applications.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Novel RNA transcript signatures for accurate age prediction using RNA profiling

  • 期刊论文

  • 发表日期：

    2026年4月24日

查看全文
  • Anqi Chen,
  • Jinyuan Zhao,
  • Fan Yang,
  • Yanan Liu,
  • Yujia Xuan,
  • Mengxiao Liao,
  • Yanfang Lu,
  • Yu Xing,
  • Suhua Zhang,
  • Chengtao Li

摘要

With advances in transcriptomics, molecular markers identified at the RNA level are being widely explored in age prediction. RNA-seq reads allow for reliable quantification of transcript-level expression by next-generation sequencing (NGS) platforms. Given that post-transcriptional regulation plays a critical role in aging, transcript isoforms and their expression may serve as novel biomarkers for age prediction. In this study, a total of 171,838 transcripts were identified from the transcriptome expression profiles of blood samples from 127 Han Chinese individuals. Differential expression analysis identified 590 transcripts with age-related expression changes across different age groups, and 16 age-related transcripts were identified by Spearman correlation analysis. Four age prediction models were constructed using these transcripts, and their performance was evaluated with an independent test set comprising 150 samples. Among these models, the Extreme Gradient Boosting (XGBoost) model exhibited the optimal performance, with a mean absolute error (MAE) of 7.609 years. Furthermore, explainable Shapley Additive exPlanations (SHAP) analysis was performed on this optimal XGBoost model, which identified four transcripts with negligible contributions to model performance. Ultimately, a final refined age prediction model constructed from the remaining 12 transcripts achieved an MAE of 7.714 years. Overall, this study investigated the potential of transcript expression levels in age prediction, providing new biomarkers and methodological insights for forensic applications.