Research On The Reconstruction Method Of Omnigenic Network Based On The Variable Selection And The Ordinary Differential Equations

Biological complex traits are a type of traits controlled by multiple genes and influenced by environmental impressions.Traditional genome-wide association study(GWAS)can identify significant loci associated with complex traits in genetic mapping studies.However,this method is unable to determine the interaction between different significant loci during the trait development process,so the genetic structure of complex traits is not fully understood.With the development of systems biology,gene regulatory network(GRN)has become an effective tool for studying complex traits,especially the omnigenic network model,which provides a theoretical basis for us to include all genes into the study at the same time to explain the complete heritability of complex traits.However,the omnigenic network model is still in the theoretical stage.Many existing methods of gene network construction have their own characteristics,but in the face of ultra-high-dimensional genetic data and multi-time dynamic phenotypic data,these existing methods have their own shortcomings in the practice of omnigenic network construction.To solve this problem,we developed a system of ordinary differential equations(ODEs)for describing the complete inter-regulation network relationship of all genes in the process of the occurrence and development of quantitative traits.We used the variable selection algorithm to reduce the dimension of each ODE and deduced the algorithm of parameter estimation to get the quantitative interaction relationship between nodes in the ODEs system.And then,a sparse,directed,and weighted omnigenic network that regulates the development of quantitative traits was reconstructed.On this basis,this study further applies the network construction method to the data obtained from the Arabidopsis thaliana growth experiment:1)Comparing the calculated total gene regulation network relationship of Arabidopsis thaliana with the existing data in the KEGG PATHWAY database,it was found that the partial accuracy rate of the network results that could be verified in the database reached 90.21%;2)The determination coefficient of the fitting effect of each differential equation was calculated,and the determination coefficient of the differential equation above 0.9 accounted for 92.6% of the total;3)A comprehensive computer simulation analysis was conducted for the proposed wholegene network construction method,and it was found that the accuracy of the two key steps in the method could reach 0.931 and 0.879 respectively under general practice conditions.Then,arabidopsis thaliana GO database was used to annotate the omnigenic network obtained,and the gene modules that played a key role in the process of stem height growth were.Key genes that raise the stem height growth module is GO: 0080179,GO: 0010282,GO: 0004846,etc.,cut role module is the key genes GO: 2001143,GO: 0044599,GO: 0052722,etc.Finally,this study analyzed the information in the omnigenic network of Arabidopsis thaliana stem height from three perspectives: 1)The effects of a single SNP on the whole gene network were analyzed,and the key nodes that played a single specific function in the network were extracted,to locate the key genes AT1G37405 and AT2G47550 that played up-regulated and down-regulated roles in the stem height growth of Arabidopsis thaliana,and the key genes AT3G55400 and AT3G13750 that played up-regulated and down-regulated roles;2)The functional relationships between SNPs and SNPs were analyzed,and the SNPs in the whole gene network were divided into 8 types according to their respective functional relationships,and representative key genes of the above 8 types were identified;3)The functional effects of SNPs on the growth of Arabidopsis thaliana were analyzed by clustering,and the functional types of SNPs on the growth of stem height were inferred from the results.To explore the relationship between the network structure and the actual physiological function of genes,the above sites with special structural characteristics were annotated in detail.The effectiveness and accuracy of the proposed method are demonstrated by database comparison and computer simulation;The analysis of the actual data shows that this method is feasible and can provide abundant and clear gene interaction information for the study of complex traits.It shows that the new method developed by this research provides a powerful tool for in-depth analysis of the genetic structure of complex traits,and provides a new method and ideas for studying genes with unknown functions using the whole gene regulatory network.