Scholarlink

The rapid advancement of high-throughput sequencing technologies has facilitated extensive research into RNA modifications, particularly N1-methyladenosine (m1A), which plays a critical role in RNA stability, translation, and cellular functions. However, accurately identifying m1A modification sites from large-scale RNA datasets remains challenging. To address this, we propose SeqAlignXGBoost, a computational framework that integrates advanced feature extraction techniques—including k-mer features, Position Specific k-mer Nucleotide Composition (PseKNC), Pseudo-Discrete Nucleotide Composition (PseDNC), and sequence alignment features—with the XGBoost algorithm for efficient m1A site prediction. The model employs incremental feature selection to optimize performance and reduce computational complexity. Experimental results demonstrate that SeqAlignXGBoost outperforms traditional machine learning models, achieving a balanced performance. SeqAlignXGBoost provides a robust and scalable solution for m1A modification site identification, advancing RNA modification research in genomics and bioinformatics.

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

SeqAlignXGBoost: Sequence Alignment and Feature Selection for m1A Modification Site Identification

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Yizheng Wang,
  • Yijie Ding,
  • Quan Zou

摘要

The rapid advancement of high-throughput sequencing technologies has facilitated extensive research into RNA modifications, particularly N1-methyladenosine (m1A), which plays a critical role in RNA stability, translation, and cellular functions. However, accurately identifying m1A modification sites from large-scale RNA datasets remains challenging. To address this, we propose SeqAlignXGBoost, a computational framework that integrates advanced feature extraction techniques—including k-mer features, Position Specific k-mer Nucleotide Composition (PseKNC), Pseudo-Discrete Nucleotide Composition (PseDNC), and sequence alignment features—with the XGBoost algorithm for efficient m1A site prediction. The model employs incremental feature selection to optimize performance and reduce computational complexity. Experimental results demonstrate that SeqAlignXGBoost outperforms traditional machine learning models, achieving a balanced performance. SeqAlignXGBoost provides a robust and scalable solution for m1A modification site identification, advancing RNA modification research in genomics and bioinformatics.