| 引用本文: | 谭超俊,顾万君,谢雪英.环状RNA计算预测方法的研究进展[J].生物信息学,2021,19(2):75-84. |
| TAN Chaojun,GU Wanjun,XIE Xueying.A review on prediction methods of circular RNAs[J].Chinese Journal of Bioinformatics,2021,19(2):75-84. |
|
| 摘要: |
| 环状RNA(circular RNA,circRNA)是一类具有重要生物作用的内源性RNA,大多在可变剪接过程中通过5’端和3’端反向共价连接形成闭合环状结构。目前,环状RNA的识别策略主要分为两大类:一类方法从高通量测序(RNA-seq)数据中检测反向剪接位点,另一类直接从RNA序列中检测成环特征。由于数据本身和识别方法的不足,依赖高通量测序数据的识别工具存在假阳性率高和不同工具间重合率低等缺点。因此,充分利用序列本身的特征来识别环状RNA是环状RNA识别的研究方向。本文总结了8种基于序列特征预测环状RNA的工具,并给出它们在测试数据集上的测试结果,为后续研究和优化提供数据支持。 |
| 关键词: 环状RNA 成环预测 深度学习 |
| DOI：10.12113/202004004 |
| 分类号:Q522+.6 |
| 文献标识码:A |
| 基金项目:国家自然科学基金项目(No.61372164;No.61471112;No.61571109). |
|
| A review on prediction methods of circular RNAs |
|
TAN Chaojun, GU Wanjun, XIE Xueying
|
|
(School of Biological Science and Medical Engineering, Southeast University, Nanjing 210096, China)
|
| Abstract: |
| Circular RNA (circRNA) is a type of endogenous RNA, which is in covalently closed form by back-splicing of 5′ and 3′ ends. Two strategies are used in circRNA prediction. One is to detect back-splicing sites from RNA-seq data, and the other is to predict circRNAs from sequence features using machine learning methods. Due to the high false-positive rate of detection tools using RNA-seq data, there are low overlaps between different sequencing-based detection tools. Therefore, tools making full use of the characteristics of sequence itself are becoming the major research direction in circRNA prediction. Here, we reviewed eight prediction tools using sequence features, and summarized their performance on the test dataset. |
| Key words: circRNA Cyclization Deep learning |
