New Methods For Analyzing MiRNA Target Genes And The Research Of Its Applications

1. Genome analysis aided prediction of miRNAs and their targetsIntroduction miRNAs play roles via their targets. Investigation of the functions of miRNAs relies on the ability to identify the mRNA targets they control. Objective Predicting inflammation associated miRNAs and their target genes. Method Combining genome analysis results with miRNA target prediction. Result Important candidate miRNAs and targets are identified by combining analysis and therefore can be used for further experimental investigation. Conclusion With the aid of genome analysis, false positive results of miRNA target prediction can be reduced.2. Measure gene(gene cluster) functional similarities based on GO annotation and the implementation of GOSemSim packageIntroduction GO provide computation accessible gene annotation data. Objective Measure gene(gene cluster) functional similarities. Method Five methods based on information content and GO graph structure respectively. Result Implement an R package called GOSemSim which calculate gene(gene cluster) functional similarities based on GO annotation. Conclusion GOSemSim implement five methods and support five species.3. Clustering human miRNA based on their targets' functionsIntroduction Cluster analysis was used to analyze high-throughput genomic data. cluster analysis can be used in extract hidden data structure, future data prediction and classification. Objective Functional clustering all human miRNA. Method Used GOSemSim package to cluster miRNA based on their targets' functional annotation. Result All human miRNA can be divided to 28 sub-cluster by kmeans algorithm. Conclusion Cluster structure indicate functional similarities between miRNA. The more closer two miRNA in cluster tree, the more functional similarity between them.