| Summary: | 碩士 === 亞洲大學 === 生物資訊學系碩士班 === 98 === Abstract
MicroRNAs (miRNAs) gene prediction is an important area of research in computational biology. The whole human genome has a size of 3 billion base pairs which make the prediction formidable. This thesis focus on developing a miRNA gene cluster prediction tool based on the secondary structures and sequence similarity of miRNAs. Prediction performance of the tool is analyzed by comparing with the miRBase dataset. We applied the prediction tool to 23 chromosomes as a test which consists of 74 miRNA gene clusters. The sensitive specificity and F1 measure are 62.2%, 40.2% and 48.8% respectively. Both of the miRNA gene prediction tools, ProMirII-g and miR-abela, achieve a better sensitivity, nevertheless, miRGCT achieves a better specificity and F1 measure. The difference is mainly due to the fact that the mature miRNA sequences used in miRGCT are relative fewer, i.e. year 2007 records, which can possibly degrade our prediction accuracy.
|
|---|
