Research On Phenotype Correlation Analysis Algorithm Based On Gene Association Network

The Genome sequencing opens the door to "group" research,in which the relationship between genotypic variables and phenotypic variables is an important research subject.The clarification of the relationship between phenotypic molecular origin and phenotype has become a key subject in bioinformatics and medical research.From the biological point of view,the phenotype will undergo a complex process to occur and develop,this process is accompanied by multi-gene interactions.With the recent experimental level and detection methods' rapid development,the completeness of genetic data is able to be improved,thus,the research methods based on gene networks provide new ideas to analysis phenotypic occurrence and development.Through the integration of all kinds of bioinformatics knowledge,a gene network was constructed to study the phenotypic correlation analysis algorithm,so as to solve the phenotype analysis and prediction problems.In my study,the physical protein interaction group data and gene phenotype data were integrated to construct gene network,and a phenotype correlation algorithm was introduced to analyze the relationship between phenotypes in gene network and its performance was verified by experiments.The main research contents are as follows: The gene network was constructed by integrating the interaction groups of physical proteins.This study associate phenotype and gene data to identify and locate phenotypic modules in the gene network and introduce the self-defined variable in the network to identify the polymerized gene modules,and carry out statistical analysis.We propose and improve the algorithm for phenotypic similarity analysis.Based on the theory of phenotype similarity and the interaction of shared proteins,a phenotypic similarity algorithm based on network separation and neighbor frequency algorithm of gene was proposed.We use the mature algorithms and data sets to verify the results,and use the mature data sets including the gene ontology and corresponding similarity algorithms to verify the performance of our algorithm.Also,the random walk with restart algorithm was introduced to compare the phenotypic similarity.In my study,the phenotypic correlation analysis algorithm proposed in the gene network can well quantify the relationship between phenotypes,and this study summed up the completeness of the genetic network,the initial weight of the network and other factors on the impact of different algorithms to provide help and guidance to further explore the phenotype of the occurrence and development of mechanisms.