Vicinal: a method for the determination of ncRNA ends using chimeric reads from RNA-seq experiments

Non-coding (nc)RNAs are important structural and regulatory molecules. Accurate determination of the primary sequence and secondary structure of ncRNAs is important for understanding their functions. During cDNA synthesis, RNA 3′ end stem-loops can self-prime reverse transcription, creating RNA–cDNA chimeras. We found that chimeric RNA–cDNA fragments can also be detected at 5′ end stem-loops, although at much lower frequency. Using the Gubler–Hoffman method, both types of chimeric fragments can be converted to cDNA during library construction, and they are readily detectable in high-throughput RNA sequencing (RNA-seq) experiments. Here, we show that these chimeric reads contain valuable information about the boundaries of ncRNAs. We developed a bioinformatic method, called Vicinal, to precisely map the ends of numerous fruitfly, mouse and human ncRNAs. Using this method, we analyzed chimeric reads from over 100 RNA-seq datasets, the results of which we make available for users to find RNAs of interest. In summary, we show that Vicinal is a useful tool for determination of the precise boundaries of uncharacterized ncRNAs, facilitating further structure/function studies.