An Automotion Software And WEB System For Predicting & Searching Gallus MicroRNA Precursor

MicroRNAs (miRNAs) are short ribonucleic acid (RNA) molecules, on average only 22 nucleotides long and are found in all eukaryotic cells, except fungi, algae, and marine plants. miRNAs are post-tRNAscriptional regulators that bind to complementary sequences on target messenger RNA tRNAscripts (mRNAs), usually resulting in translation repression and gene silencing. The human genome may encode over 1000 miRNAs, which may target about 60% of mammalian genes and are abundant in many human cell types. miRNA guides RISC to direct regulation of protein-coding mRNAs depending on the complementarity between the miRNA and its target mRNA. Recent studies proved that miRNAs were stongly associated with genetic diseases including oncogenesis and could serve as therapeutic targets or agents for different types of cancers.Chicken (Gallus) is an important animal model in scientific research and is of economic value worldwide for meat and egg production. Although it may be essential to identify all Gallus miRNAs to obtain better insight into the biological function, the current release of miRBase17.0 contains only 499 Gallus miRNA sequences while 1424 pre-microRNAS, 1733 mature microRNAs of human has been listed in miRBase 17.0. Identification of miRNAs has relied on two predicted strategies: One computational strategy based primarily on comparison to pre-miRNA hairpin structures, which relied on the sequence conservation of miRNAs across related species. The other computational strategy based on features of local contiguous structure-sequence information, which relied on these features to classify real and pseudo pre-miRNAs. Here, we developed a software package, called GluMiRFilter, to filter the pre-miRNA candidates according to sequence and structure characters of pre-miRNAs. Then we developed a novel classifier based on support vector machine for classification of real and pseudo pre-miRNAs, called GluMiRPred, to predict Gallus pre-miRNAs. Only a few of Gallus miRNAs in miRBase were computationally predicted on the basis of sequence homology to known miRNAs from other sepecies. GluMiRPred relied on SVM and could predict Gallus species-pecific pre-miRNAs. Trained on 300 Gallus pre-miRNAs and 125 pseudo Gallus hairpings, GluMiRPred achieves accuracy of 90.3%. We also tested it with ROC cure and the result is 0.93, which means that the accuracy of this software is 93% measured with ROC. We tested GluMiRPred on the remaining 199 Gallus pre-miRNAs and 1000 pseudo hairpins, and its result reports sensitivity of 100%, specificity of 88.4% and accuracy of 90.3%. These results proved that GluMiRPred was an effective ab initio approach for predicting Gallus pre-miRNAs.We filtered chiken miRNAs with GluMiFilter and GluMiPred and got 7471 pre-miRNAs.