基于双向门控循环单元的5-甲基胞嘧啶位点预测  

Prediction of 5-methylcytosine site based on BiGRU

作　　者：黄修威 方中纯 李海荣 HUANG Xiuwei;FANG Zhongchun;LI Hairong(College of Information Engineering,Inner Mongolia University of Science and Technology,Baotou 014010,China;Engineering Training Center(College of Innovation and Entrepreneurship Education),Inner Mongolia University of Science and Technology,Baotou 014010,China)

机构地区：[1]内蒙古科技大学信息工程学院,包头014010 [2]内蒙古科技大学工程训练中心(创新创业教育学院),包头014010

出　　处：《中南民族大学学报（自然科学版）》2023年第6期768-774,共7页Journal of South-Central University for Nationalities：Natural Science Edition

基　　金：内蒙古自治区自然科学基金资助项目(2020MS06028);内蒙古自治区高等学校科学研究项目资助项目(NJZY21387);2020年教育部产学合作协同育人项目资助项目(202002107034)。

摘　　要：5-甲基胞嘧啶(5-methylcytosine,m5C)是一种重要的转录后修饰,大量证据表明,m5C在许多生物学过程中起着至关重要的作用.准确鉴定m5C位点有助于更好地了解其生物学功能.为此提出了一个名为pm5C-BGRU的模型,该模型通过拼接独热编码(One-hot encoding)和核苷酸化学性质(nucleotide chemical property,NCP)进而对RNA序列进行特征提取,并基于双向门控循环单元(Bidirectional Gated Recurrent Unit,BiGRU)来识别m5C位点.将该方法在人类、小鼠和拟南芥三个物种的m5C数据集上进行建模和测试,并对照已有的预测模型进行评估.结果表明,pm5C-BGRU在交叉验证和独立数据集测试中均取得优异效果,该模型有望成为鉴定m5C位点的有力工具.5-methylcytosine(m5C)is an important post transcriptional modification.More and more evidence shows that m5C plays an important role in many biological processes.Accurate identification of m5C site is helpful to better understand its biological function.A model named pm5C-BGRU is proposed.The model extracts the features of RNA sequences by splicing One-hot encoding and nucleotide chemical properties(NCP),and identifies m5C sites based on the bidirectional Gated Recurrent Unit(BiGRU).The method was modeled and tested on m5C datasets of Homo sapiens,Mus musculus and Arabidopsis thaliana,and evaluated against the existing prediction models.The results show that pm5CBGRU has achieved excellent results in cross validation and independent testing,and the proposed model is expected to be a powerful tool for identifying m5C sites.

关 键 词：5-甲基胞嘧啶 序列编码 双向门控循环单元 预测 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]