Analysing Mirna Interaction And Conservative Sites Based On Secondary Structure

With the accomplishment of human genome project, biological data has experienced a rapid growth. Traditional experimental methods show limitations to deal with large data sets. Increasing bioinformatics methods have been applied to biological data research. How to mine implicit biological association with efficiency and accuracy becomes a key issue. In recent years, some researches show miRNAs play an important role in gene regulation. More and more studies are used to miRNA research, like miRNA prediction, target prediction, secondary structure prediction, and so on.In this thesis, it reviews the previous research of miRNA modeling and summaries the advantage and disadvantage of these methods. Two methods to model miRNA secondary structure are proposed. The one method models miRNA in term of stem and loop. It focuses on the number of each stem and loop. The next considers the category of base and structure in each site.To overcome the limitations of the previous study on miRNA relationships, structure similarity in combination with joint entropy are employed to explore the relationship of miRNAs. RMS (Root Mean Square) is applied to measure the distance of two miRNAs and the similarity measure is to evaluate their similarity. Joint entropy is used to measure the importance of related miRNAs. The related miRNAs that have higher joint entropy are viewed as more important. The experiment is conducted by using Drosophila melanogaster and Anopheles gambiae data set collected from miRbase database. The result demonstrates our method is able to correctly identify the related miRNAs and unknown interacted miRNAs, which assist in correct and effective drug design.For conquering the shortage of mature miRNA site research, information entropy is employed to discover mature miRNA conservative sites. They contain single nucleotide information entropy and adjacent nucleotide information entropy. The site with lower information entropy is regarded as more conservative. In other word, this site is more important with respect to miRNA function. In addition to information entropy, secondary structure is taken into account in this thesis. Thus, it ensures that the discovered sites by our approach are important from biological point of view. The experiments demonstrate that our methods not only verify some known important sites, and discover new unknown site. This is useful in understanding evolution and biological function of miRNAs.