Research On The Construction Algorithm Of The Key Factor Network For Magnaporthe Oryzae Oryza Sativa L.Interaction Based On Multi-omics

Oryza sativa is one of the most essential food and cash crop,directly impacting agricultural development and food security with its high economic value.However,rice blast caused by Magnaporthe oryzae can significantly reduce yield and degrade the quality of rice.To develop effective and comprehensive control strategies,current research must analyze the mechanism of fungal pathogen infestation on host plants.Traditional molecular marker studies based on single omics are insufficient in revealing the complex biological regulatory processes involved.Integrating vertical omics can provide a more comprehensive understanding of the internal mechanisms of organisms.However,studying the interaction mechanism across species is hindered by numerous conditions,and investigating multi-omics complex systems through biological experiments alone is difficult.Machine learning makes it possible to integrate multiple omics,but the limitations of plant experiments make it impossible to connect each omics with unique annotation information.This difficulty in standardizing different omics networks poses challenges to multi-omics studies of fungi-infected plants.This research proposed a multi-omics integration method based on heterogeneous node relationship prediction to solve the problem of heterogeneous association in the multi-omics studies of fungal plants.Based on the experimental research and data of fungal plant interaction,this method constructs gene co-expression network by WGCNA(Weighted Gene Co-Expression Network Analysis),a transcriptome network of rice m RNA targeted by mi RNA of Magnaporthe oryzae,and the same-species and cross-substance protein network by String and Interolog methods.Then the local network structure of node among single omics network was obtained by deep random walk.Meanwhile,the link features of heterogeneous nodes are constructed.Then this research improved DBSCAN(Density-Based Spatial Clustering of Applications with Noise)algorithm based on convolutional feature extraction so that the algorithm can achieve the high-dimensional feature clustering ability.This method predicted 191 potential heterogeneous network links,and a multi-omics hierarchical heterogeneous network of rice interaction was constructed.At the same time,281 key factors were screened,including 60 key pathogenic factors and 221 rice disease resistance factors.Through GO enrichment analysis and KEGG metabolic pathway analysis of key factors,this paper revealed the biological processes involving pathogenicity and resistance factors in plant interaction of Magnaporthe oryzae.It verified the regulatory effects of some enrichment results under biological stress through experimental results in the articles.This paper demonstrates that the proposed multi-omics integration algorithm effectively constructs the multi-omics network of Magnaporthe oryzae Oryza sativa interaction,identifying key factors.This approach successfully addresses the problem of heterogeneous association among different layer networks in multi-omics studies of fungal plants.Moreover,the identified key factors offer a collection of molecular markers for biologists to test,and this research lays a solid foundation for further studies on the pathological attack of plants by fungi.These findings provide an essential direction for studying the mechanism of rice blast infestation and formulating effective rice blast control strategies.