Life Science Knowledge Network System Building And Network Information Analysis

With the high-throughput data analysis produced a large number of data, bioinformatics databases and systems biology in life science research are increasingly important. A large number of databases and web services make the biologists facing the danger of lost in data, in addition to how to effectively organize and use information has become the focus of bioinformatics. In order to build a unified Bioinformatics framework for effective organization, unity and analysis of these different sources and types of data, we analysesd the basis of biological information data structure and information constitute. In the processing of raw data this study, we designed a data framework based on the concept node and relationship among them. We also built an ontology dataset for massive life science concepts, a new semantic-based literature search engine. We have also developed a new network analysis algorithms, combined with the standardized information scores, we can calculate and sort the most relevant concepts, find out information path which eventually product a possible biological explanation. Based on the basic research and data processing, we have developed a life science knowledge engine called BioPubInfo(http://www.BioPubInfo.org), including literatures search engine and knowledge analysis engine. Knowledge of network analysis engine has completed initial development of the interface and back-end settings, the literature search engines still on development. We worked out a set of design and analysis algorithm to handle mass data of life sciences, and process massive amounts of data using the network structure to predict new knowledge. The new algorithms take advantage of graphic database and graphical data structure frame, can calculate one hundred million level concepts in real-time. Using this tool, we analysed the human diseases and Arabidopsis, rice phenotype candidate genes.Based on network, we predicted the relationship between rice phenotypes and genes, integrated with other information and established QTXtoGene integrated analysied platform which will add more species and traits.We also analyzed the expression data in Arabidopsis under salt stress, built a transcript regulatory network of Arabidopsis roots under salt stress at different times. Using new horizontal gene transfer(HGT) methods analysis and found 10 candidate HGTs in silkworm genome. Mutual information method was used for analysis the residues evolution network in influenza virus receptors.