As an essential aspect of structural biology, RNA secondary structures play a crucial role in maintaining molecular stability and regulating various biological functions. Since experimental approaches for resolving RNA secondary structures are often complex and costly, developing efficient and accurate prediction methods has become a key direction in RNA structural research. Computational biology, an important analytical tool, provides structural biology with theoretical foundations and algorithmic support. Due to its great potential in RNA secondary structure prediction, it has attracted extensive attention in biomedical research. This article first outlined the main computational methods for RNA secondary structure prediction, including energy-based methods, multiple-sequence methods, traditional machine learning methods, deep learning methods, and tertiary structure-based RNA secondary structure prediction methods and compared the advantages and disadvantages of various algorithms. Then, this article discussed the applications of related techniques in biomedical fields, with a particular focus on the identification of RNA-binding protein sites. Finally, it provided an outlook on the future development of RNA secondary structure prediction methods. This review is expected to provide important references for relevant research.
RNA二级结构作为结构生物学研究的重要内容，在维持分子稳定性和调控多种生物学功能中发挥重要作用。由于通过实验方法解析RNA二级结构的过程复杂且成本较高，发展高效准确的预测方法已成为RNA二级结构研究的重要方向。计算生物学作为重要分析工具，为结构生物学提供了理论基础和算法支持，在预测RNA二级结构方面展现出巨大潜力，在生物医学相关领域得到了广泛关注。本文首先综述了RNA二级结构预测的主要计算方法，包括能量方法、多序列方法、传统机器学习方法、深度学习方法和基于三级结构的RNA二级结构预测方法，并对比分析了各类算法的优缺点;然后探讨了相关技术在生物医学领域中的应用，重点对RNA结合蛋白位点识别进行了总结;最后对RNA二级结构预测方法的发展前景进行了展望。本文可为相关研究提供重要参考。.
Keywords: RNA secondary structure; RNA-binding proteins; computational biology; prediction; structural biology.