潜在功能性长非编码RNA预测算法研究及应用

Long non-coding RNA (lncRNA) has important biological functions. With the development of high-throughput sequencing technology, a large number of lncRNAs have been obtained through the prediction algorithm, but there are only more than a thousand experimentally validated functional lncRNAs. It is controversial whether all the transcribed lncRNAs are functional. Using traditional experimental methods to verify the function of lncRNAs is time consuming, so it is very important to develop the algorithm to predict potential functional lncRNA. Therefore, this project mainly focuses on the following studies: (1) On the basis of the EVLncRNAs database, which is the database we built about experimentally validated lncRNAs manually extracted from literature, this project will use experimentally validated lncRNA as positive set, lncRNAs predicted from high-throughput data and mRNA as negative set, then mine structural and experimental features, combined with the feature from sequence, to develop potential functional lncRNA prediction algorithm, which could identify reliable and functional lncRNA more accurately, quickly and efficiently; (2) The prediction algorithm will be developed as an online prediction platform for researchers to use; (3) Using our algorithm, the new predicted functional lncRNAs will be found, and the function of lncRNAs will further be validated by experiments.

长非编码RNA（lncRNA）具有重要生物学功能。随着高通量测序技术的发展，通过预测算法得到海量lncRNA，但实验验证有功能的lncRNA仅有千余条，是否转录的lncRNA均具有功能存在争议。然而通过传统实验方法验证lncRNA功能耗时耗力，所以发展潜在功能性lncRNA预测算法具有重要意义。本项目研究重点：（1）创新思路，利用我们建立的收录所有文献报道的实验验证lncRNA数据库——EVLncRNAs，用实验验证lncRNA作为正样本，高通量预测得到的lncRNA和mRNA作为负样本，挖掘结构和实验特征，结合序列特征，发展预测可靠和潜在功能性lncRNA的算法，更加准确、快速、高效地鉴别功能性lncRNA；（2）将预测算法开发为在线预测平台，方便科研人员使用；（3）应用构建的预测算法，发现新的功能性lncRNA并进行实验验证，更加深入地研究lncRNA的功能。

长非编码RNA（lncRNA）是长度大于200nt、不能翻译成蛋白质的非编码RNA，在调控基因表达和多种生物学过程中发挥着重要作用。高通量转录组测序技术的发展促进了对lncRNA及其功能的发现，但是多达几十万的高通量lncRNA中只有几千条被实验证实。本课题中，1）基于我们自己建立的实验验证lncRNA数据库EVLncRNAs，使用实验验证lncRNA区别于高通量lncRNA、mRNA的序列、结构和实验特征，发展了直接预测潜在功能性lncRNA的预测算法EVlncRNA-pred，并开发为在线预测平台。2）应用EVlncRNA-pred算法对本实验室获得的马铃薯转录组数据进行预测，发现了潜在的功能性lncRNA，正在进行实验验证。3）在完成项目原有计划的基础上，进一步收集、丰富了实验验证lncRNA数据集，并建立为更新的EVLncRNAs 2.0数据库，使用CNN和DNN深度神经网络，发展了潜在功能性lncRNA深度学习预测算法EVlncRNA-deep，为功能性lncRNA预测提供新思路，使发现新的功能性lncRNA更加高效。4）此外，分析了lncRNA在甲状腺癌、乳腺癌、肺鳞癌等癌症中的表达及与患者预后的关系，为挖掘有效、可靠的癌症生物标志物提供参考。本项目在国际高水平期刊Nucleic Acids Research、RNA Biology、Frontiers in Plant Science等发表论文6篇。