Abstract:

MicroRNA (miRNA) are a class of endogenous 21～24 nucleotide non-coding small RNAs. They may participate in a wide range of genetic regulatory pathways and play an important role in the development of both animals and plants. The number of miRNA genes in every species is still unknown. Bioinformatics analytic method has been used to find novel miRNA genes. In this study, a computational system called MiRdetector is developed, which is used to predict and identify miRNA genes. The computational system is based on homology search and prediction of stem-loop structure. Oryza sativa miRNA genes were used to examine the prediction quality. In total, 140 in 155 samples were correctly identified. The prediction accuracy reaches 90.32%. Moreover, detected 95 potential miRNA genes were detected. The false positive rate was low, so the data set of candidate miRNA genes can well be used for further verification in experiments.

Key words: biological genome, BLASTN, MiRdetector, stem-loop structure

miRNA genes