Long non-coding RNA functional annotation : machine learning approaches

Abstract

Long Non-coding RNAs (lncRNAs) play crucial roles in complex pathological and physiological processes. However, only a few of lncRNAs are well characterized. lncRNA functional annotation mainly includes two parts: lncRNA annotation and lncRNA function exploration. The biological experiments for lncRNA functional annotation are costly and time-intensive, and the characteristics of lncRNAs pose further challenges to their understandings. Therefore, in this thesis, I aim to develop machine learning approaches to explore the lncRNA functional annotation. I start by identifying the RNA transcripts from background DNA sites, then I try to distinguish the lncRNAs from coding RNAs. After that, I develop computational approaches to indicate the lncRNA functions by identifying the types of biomolecules that a lncRNA would interact with and then focusing on a certain type of interaction, i.e. DNA:lncRNA triplex, to reveal the lncRNA function. The results show that the proposed approaches are effective for lncRNA functional annotation.