Application Of Protein Interaction-based Systems Genetics Method In Plant Functional Gene Identification

Connecting functional genes to target phenotypes remain one of the research focuses in genetics currently.Based on high-throughput sequencing technology,Genome-Wide Association Study(GWAS)is used to identify the relevant gene loci of the target phenotype in a variety of plants.At present,GWAS analysis of A.thaliana,rice and maize has produced a large amount of research data and established many professional databases.However,it is still an important challenge to determine the association between genes and phenotypes due to the false positives of traditional GWAS results.Therefore,this research used systems genetics methods to study the above problems based on the summary GWAS data of plants(e.g.Arabidopsis thaliana,Oryza sativa,and Zea mays).Firstly,this work re-identified functional genes in the target phenotype based on GWAS data through Gene Rank,K-shell,Hot Net diffusion-oriented subnetworks(Hot Net2)and Knowledge Graph(KG)in the days to flowering trait,the root morphology trait and the yield trait.Moreover,the gene enrichment ratio calculated by the above multiple methods was compared,and it was verified that the method based on PPI can effectively improve the functional gene enrichment ratio.Therefore,a hybrid model was proposed to effectively improve the gene enrichment ratio of plant functional genes(up to 40%)based on A.thaliana GWAS data and PPI network topology data,and combined with Gene Rank algorithm,comprehensive scoring algorithm and Hot Net2 algorithm.Besides the hybrid model was applied in 462 phenotypes of A.thaliana.Secondly,we applied the hybrid model to the gene-phenotype association study of rice and maize in order to improve the efficiency of crop functional gene identification.After genotype imputation of rice and maize GWAS summary data,the genotype of rice and maize research has increased by several tens of times,and the relationship between phenotype and SNP in rice and maize of GWAS results is mapped to the relationship between phenotype and gene.Based on the processed gene-phenotype association data,this work successfully identified multiple gene-phenotype associations in rice and maize through the hybrid model.Many of these phenotypes also have genes that are different from the GWAS summary data,and these genes are support by literatures.Moreover,the mechanism analysis of the differential genes in target phenotypes verified the correlation between target phenotypes and differential genes.Finally,this paper uses the website architecture LAMP to design and develop a plant functional gene identification database PlantGWASRank(PGR,http://47.242.161.60/Plant/).PGR database provides: 1)retrieval,browse,analysis,visualization and download of the result data of A.thaliana,rice and maize using multiple methods(GWAS,Gene Rank,K-shell,Hot Net2,KG and the hybrid models);2)the hybrid model online calculation,background operation,and offline email sending of result data;3)software package download of multiple methods;4)help and other functions.In conclusion,due to protein interaction-based systems genetics method can improve the efficiency of plant functional gene identification,a hybrid model was proposed to improve the functional gene enrichment ratio of the plant target phenotypic based on GWAS data and PPI network topology data and combined with Gene Rank algorithm,comprehensive scoring algorithm and Hot Net2 algorithm.The hybrid model has been effectively used in A.thaliana,maize and rice,and the PGR database has been designed and developed to realize data sharing.