| Summary: | 碩士 === 國立成功大學 === 電機工程學系碩博士班 === 97 === MicroRNA are short RNA about 21~23nt which exist in non-coding region. MiRNA play important rules in post-transcriptional regulation, reveals that RNA is not only a carrier of gene information, but also a mediator of gene expression . Detecting unknown miRNAs get more attentions. Using computational method can help us find new miRNAs effectively. So how to use these computational method to detect new miRNAs is more important.
However, most advanced miRNA prediction algorithms on miRNA discovering can not process large-scale genomic sequence as input. This article presents a tool that can predict pre-miRNAs automatically form large-scale genomic sequence. Our tool isa based on a ab initio tool, miR-KDE, and combine a secondary structure filter and homology analysis to help users to find pre-miRNAs from large-scale sequence. MiR-KDE can efficiently predict pre-miRNAs (overall accuracy 94.7%). In our tool, we add secondary structure filter and homology analysis that make our can process length of 10M nt sequence and report result in reasonable time (~20sec / 100k nt). Both the limit of input sequence and process time, our tool shows batter performance then other large-scale pre-miRNA predicting tools.
|
|---|
