The MiRNA Target Gene Prediction Algorithm Based On H1N1

On April 2009, a worldwide epidemic of influenza a virus outbreak led to hundreds of thousands of infections and deaths of thousands of people, making people raise concerns for the influenza virus again. First of all, the U.S. Centre for Disease Defense revealed that it is a new porcine influenza A H1N1 influenza virus. This virus contains new gene segments, and then this new combination of genes had never appeared in human and pig. Influenza viruses have eight gene segments, respectively called HA, NP, NS, NA, PB1, PB2, PA and M, which were similar with three sources of restructuring the North American swine influenza viruses and Eurasian swine influenza virus. The present study shows that the H1N1 influenza virus M gene exists in this fragment where there is a public feature, so with the fragment of the virus to amantadine has a strong resistance to drugs, so A H1N1 virus is Similar that there is antiviral drugs amantadine resistance. Researchers also found that influenza virus virulence and power of spread is no less biological characteristics than the previous A-type virus virulence and power of spread, but for influenza virus pathogen city and power of spread till is not very clear.Currently, the non-coding RNA (miRNA) research has become important issue of the biology and bioinformatics direction. miRNA discovery as we explore the complex mechanisms of gene regulation has opened up new roads. Research shows that one-third of all human genes are subject to the regulation of miRNA, indicating that miRNA is human central part of gene regulatory networks, so researching function and mechanism of miRNA is very significant. miRNA is a class of about 21 or 22 bases (nt) long small molecular non-coding RNA, which can be formed by the length of 70nt miRNA hairpin precursor by dicer cleavage. miRNAs are widespread in animal and plant cells, which cleavage the mRNA molecule or repress the initiation of the mRNA translation by complementary pairing with the target gene,and so as to realize the regulation of gene expression, and work for the important regulation on cell differentiation, proliferation and apoptosis.We know that miRNA discovery; accurate prediction of miRNA target genes and a clear understanding of mechanism of miRNA and target genes is currently a hot issue about miRNA research. In this paper, accurate prediction of miRNA target genes were the important task to study. There are two ways of miRNA and target gene binding; one is completely complementary pair; the other is not completely complementary pair. When the miRNA and target gene is completely complementary pair, generally achieved regulatory role through the cleavage of target gene; when the miRNA and the target gene is not completely complementary pair, the target expression is inhibited for realizing the control function. For the former pairing mode, you can write a simple computer program to predict the target genes, but this match for the latter approach, a simple procedure can not meet the miRNA mechanism of interaction with the target gene, and many methods are focused on the seed region to predict the importance of target genes, but can't often reach a good prediction effect. Machine learning methods appearing, they could acquire more biological properties from combination of miRNA and target genes and then add to the existing algorithms. The results of prediction algorithm will be more accurate and effective.As the influenza virus mutates rapidly, this feature is determined by the nature of the influenza virus itself, and can not be changed, so gives the researchers great difficulties. The traditional biological methods it is difficult to play therapy and prevention effects, and furthermore a vast majority of the target sequence of influenza virus common exists in the coding region, and is highly conservative, so the target sequence itself is a part of translation proteins. Absence or degradation of the target sequence may change nature of virus. This indicates that people should learn from the bioinformatics perspective, by the intelligent algorithm to mine in a flood and constantly updated data, finding the required information, rules, more features and miRNA and its target gene which may control influenza virus in the human body, finally reveal the importance of the pathogen city and the size of influenza virus on the level of the genome. This paper describes some important miRNA target gene prediction algorithm, and using software evaluation parameters (sensitivity, false positive rate, signal to noise ratio), and compared the difference of prediction algorithm, using TarBase, GenBank and miRBase database resources to download the influenza H1N1 virus genes, the human miRNAs and a set of target gene sequences, and the target gene data set is categorized into three data sets. Through the three data sets, selecting the required human miRNAs, and then searching H1N1 influenza virus genes by existing software, and the prediction results is drown into a table, analyzes the performance of each software with evaluation indexes. Finally, conclude advantages and disadvantages of each prediction algorithm. These conclusions are useful for the development of new target gene prediction software.It is seen from the data obtained in this study that each of the software for the prediction of new species are not particularly ideal. Further development of prediction software needs the support of new experimental techniques. With the increasing of the number of target genes verified, we will gain more statistical specimens that will provide us with more characteristic parameters thus promoting the development of gene prediction software. Meanwhile, the development of prediction software also facilitated the updates of experimental technology, technical innovation can clarify the biological mechanism between miRNA and target genes which maybe make people learn regulatory network of miRNAs clearly. The present prediction software is not particularly accurate, mainly because the complexity of mechanisms of interaction between the miRNA and target genes have been far beyond our imagination, there are still many biological characteristics not found. So more and more target genes are checked through experiments that there will be more interactive characteristics between miRNA and target gene that are used in the target gene prediction, and the calculation method will be further optimized.This paper maybe provide a theoretical basis for future researchers, while data obtained from the comparison may be made of by pure biology researchers to overcome shortcomings of the traditional method for influenza virus on the prevention and early diagnosis, which will guide the future development of the drug target genes , but also has high theoretical and practical value.