Based On SVM Of MicroRNA Multi-feature Fusion Model Research

microRNA (miRNA) is present in the genome of a class of endogenous non-codingsmall molecule protein RNA, the length of about 21 ~ 23 nucleotide mature bodylength ofabout18~29,mature,anditsexpressionistissueandtimespecific,Andsomeofthegenesin evolution is highly conservative. It is a combination with the target gene can degrademRNA or inhibit its translation and thus play a role in regulating gene expression insingle-stranded small molecule RNA, usually in the gene or between the intron, with thetarget gene mRNA3 'UTR Region (UTR) of target sites of regional mutual combination offast and effective degradation of mRNA, or inhibit the protein's usage would proteinrequired for controlling the activities of life or the best low level of miRNA involved,including early development, cell apoptosis, cell proliferation, fat metabolism, cell death,celldifferentiationandaseriesofimportantprocessinthelifecourse,thereisresearchthat,miRNAandcanceralsohaveanumberofpotentialbetweenthecontacts.Known in the mature miRNAlength between most of the 22nt ~ 24nt, and most withU at the beginning, not to C at the beginning, but its precursors are generally shortersequence, that is, miRNAprecursor hairpin loop structure that contains and shorter lengthof the sequence. In miRNA production process, the mature body with a hairpin structurefrom the precursor after Drosha processing enzyme Dicer enzyme and the formation of theshear process of the selected sites, namely, 3 'end and 5 'end positions and the formation ofthe existence of the length of the stem zone also has a certain regularity, this ruledeterminesthemiRNAmatureproteins.miRNA prediction in the process will usually give priority according to theircharacteristics with the conservative use of its sequence and structural features, namely:sequencelength,thesecondarystructureofthehairpinnumberandsizeofthedataanalysis.Since the database is now included in the data still possible false positive, and someflanking sequence of the primary transcripts as pre-miRNA is also stored in them.Therefore, miRNA feature extraction, we will first of all through the selection, screeningfrom the miRBase miRNAin 700 positive data, to ensure the accuracy of data sources, toensure that all miRNA data are identified through biochemical experiments or by a highcredibility of the verification algorithm. Followed by extraction of its characteristic sequence features and structural features intotwo parts integration,through asingletypeofSVM algorithm for the calculation of the final features of the results obtained. Using thisdata on the extraction of miRNAprecursors and the feature fusion of this process, we canimprove the body's mature miRNAprediction accuracy, in order to laythe foundation afterthemiRNAprediction.This paper, we mainly focused on miRNA prediction process for the many types ofbiological features extraction and SVM based algorithm for a single category will beeffectiveintegrationofthesecharacteristics, andcalculatethefusionresult. Forecast forthecurrent mass miRNAgenetic data, molecular biologyexperimental method alone to predictand analyze the relevant information has been difficult miRNA to meet current demand,therefore, the development of more efficient and reliable method to uncover usefulinformation miRNAsequence.the current study of small molecule RNAto solve an urgentneed for one of the problems, but also on biological research and disease treatment are ofgreatsignificance.