Risk Assessment Of Avian Influenza Transmission Based On Spatiotemporal Data Analysis

The spatiotemporal spread of avian influenza viruses has caused a serious public health problem all over the world.The corresponding studies on the outbreak of avian influenza and its spatiotemporal transmission dynamics has become a hot research issue in many fields such as epidemiology,molecular biology and phylogeny.However,previous research has limitations for this complicate problem.First,there exist many complicated factors for the underlying mechanism,while some previous studies only considered the geographical properties of viral sequence,without focusing on the biological properties of virus hosts such as spatiotemporal movement.Second,it is difficult to accurately quantify the spatiotemporal propagation patterns of viruses solely based on the phylogenetic analysis at the microscopic level.At the macroscopic level,due to the limitations of data scale and data collection methods for the spatiotemporal data of virus hosts,current studies have difficulty to accurately describe the spatiotemporal spread of avian influenza viruses and the spatiotemporal pattern of the host at a fine-grained scale.In this paper,we focus on the outbreak of new avian influenza(H7N9)virus and its risk assessment in spatiotemporal in China and tackle the abovementioned issues.We collect spatiotemporal data from multiple sources and different disciplines,and by integration of the data mining algorithms,epidemiology,phylogeny,host space biology and other interdisciplinary methods,we systematically study the evolution and transmission mechanism of avian influenza virus.From the perspective of data collection and analysis,we design and develop a bird spatiotemporal data collection platfonn based on citizen science.Further,based on the collected bird observation data,we propose an improved hierarchical DBSCAN clustering(IHDBSCAN)algorithm and a kernel density estimation method,to analyze the biological characteristics of bird hosts and to explore the spatiotemporal distribution of different bird species and their migration patterns.Combining the bird observation data and reported case of avian influenza,we analyze their spatiotemporal distribution and migration patterns in China.At the microscopic level,we conduct phylogenetic analysis on the genetic sequences of avian influenza(H7N9)virus,and investigate the evolutionary mechanism of the virus and its spatiotemporal propagation patterns.We find that there exist three major epidemic waves of avian influenza(H7N9)virus in Southeast China,and observed that most sequences in the same municipality/provinces and in the same year are clustered together in the phylogenetic tree.At the macroscopic level,we use the Cross-Correlation Function to systematically explore the correlation between avian influenza(H7N9)virus and the spatiotemporal distribution of bird species.We conduct experiments to identify possible migratory bird species that are relevant to the avian influenza(H7N9)infections in Shanghai,Jiangsu,Zhejiang,Fujian,Jiangxi,and Guangdong provinces in China,and combine the spatial biological characteristics of different bird species to quantify the hotspots of avian influenza(H7N9)virus.In addition,we adopt the Granger Causality method to identify the causal relationship between key bird species distribution and avian influenza(H7N9)virus in six municipalities/provinces.In summary,this paper first proposes a hierarchical clustering method and a kernel density estimation method to explore the spatiotemporal distribution of migratory birds and their migration patterns in China.Then,this paper uses phylogenetic analysis methods to reveal the evolution and geographic patterns of influenza A(H7N9)infections.The experiments found that there exist three epidemic waves along the East Asian-Australian flyway,while cases in the same or nearby municipalities/provinces are clustered together with small evolutionary differences in the phylogenetic tree.In addition,this paper uses cross-correlation function and causality analysis method to identify a list of possible migratory bird species that are relevant to the influenza A(H7N9)infections in Shanghai,Jiangsu,Zhejiang,Fujian,Jiangxi and Guangdong in China.Based on the spatiotemporal distribution of identified key bird species,potential geographic hotspots are further estimated in these typical municipalities/provinces.These results and findings could provide sentinel signal and evidence for active surveillance,as well as strategic control of influenza A(H7N9)transmission in China.