Transcriptome-wide Study On The Factors Associated With LncRNA Stability In Human

Metabolism is very important in organisms,and the metabolic processes from transcript generation to degradation are of great significance.However,there exists significant differences in the stability from different transcripts.Furthermore,the stability of the transcripts is closely related to their biological functions.Abnormal expression level of transcripts,whether caused by their instability or over-stability,will be harmful to organisms.Therefore,it is very important to determine which factors affecting the stability of such a large number of transcripts.At present,the factors associated with the mRNA stability have been studied in depth,which mainly include ARE elements,RNA modification and NMD-mediated degradation pathways.However,little is known about the factors related to human long non-coding RNA(lncRNA)stability.As a type of RNA that does not encode protein,it participates in a variety of biological processes such as X chromosome silencing,gene imprinting,chromatin modification,gene transcription activation and inhibition in the cell.Therefore,deep analysis of lncRNA stability plays a key role in understanding the biological processes above.However,there are few studies on systematic analysis on the lncRNA stability in human with large-scale datasets.From the perspective of statistics,large datasets are necessary to deduce the reliable conclusions.In order to solve this problem,we took the human lung cancer cell A549 as the research object.Through transcription suppression and high-throughput sequencing,we obtained the half-life datasets including 33285 lncRNAs 24,710 mRNAs,nucleus-specific 5758 lncRNAs and 13093 mRNAs,cytoplasm-specific 1515 lncRNAs and 9833 mRNAs,and nucleus and cytoplasm-common 491 lncRNAs and2496 mRNAs.The reliability of the datasets is assured by three replicates and random sampling strategy to calculate half-life.Then,we used Spearman correlation analysis to study the relationship between the stability of lncRNA or mRNA and the number of exons,lncRNA classification,secondary structure,cellular location,mi RNA-RNA interaction,protein-RNA interaction and other factors.We got the following results.Firstly,the stability of lncRNA and mRNA is closely related to the number of exons in them.Secondly,the nuclear and cytoplasmic location of the transcript has no significant influence on the stability of the transcripts.The reason for the stability of the transcript in the nucleus lower than that in cytoplasm maybe the transcripts themselves or other factors rather than cellular location.Thirdly,protein or RNA interacting with lncRNA or mRNA will promote the stability of one-exon lncRNAs and reduce the stability of multi-exon mRNAs with a high probability.Fourth,the factors above such as transcript secondary structure affect lncRNA and mRNA stability in a non-linear manner.Finally,in order to better demonstrate the correlation between stability and various factors,we constructed a regulatory network between the stability of lncRNA or mRNA and a variety of influencing factors.It clearly demonstrated the degradation difference between the regulation of lncRNA and mRNA stability.Finally,we used the linear regression and deep learning-based nonlinear regression methods to develop lncRNA and mRNA half-life prediction models.In short,in order to study the stability of human lncRNA systematically,this study constructed the human transcriptome-wide half-life dataset,and studied the relationship between lncRNA or mRNA stability and various factors using a variety of statistical methods.These works will help researchers understand the relevant factors which affect the stability of lncRNA and mRNA,and provide a research basis for future studies on the biological metabolism of both lncRNAs and mRNAs.