Application of Data Mining in Research of Avian Influenza Virus Cross-Species Infection

Avian Influenza Virus (AIV) has already crossed species barriers to infect humans, but the reason for avian flu cross-species infection is unknown. A lot of biology experiment accumulated tens of thousands of biological information data. As a new technology based on database and statistics, data mining provides unprecedented data analysis tool to biologists and powerful means to gene and protein's analysis and extraction. In this paper, we applied feature-based clustering and classification method to the research of AIV cross-species infection, finding useful patterns, and created a web-based early warning system. We also applied entropy plot and regression analysis to discover host-associated sites in AIV cross-species infection, and got the key sites that differentiate human versus avian influenza using the same method. Compared the two sets, we expected to discover some rules about the mutation trend.