The Multi-omics Study Of Rice False Smut Pathogen Ustilaginoidea Virens Based On Bioinformatics Approaches

Rice false smut(RFS)caused by Ustilaginoidea virens is one of the most devastating rice diseases.After infecting the rice floral organs,the disease not only causes significant yield losses in rice,but also produces different types of mycotoxins,which pose severe health threats to human and animals.RFS is an emerging disease,and little is known on molecular mechanisms underlying U.virens virulence and pathogenicity.Therefore,the studies on U.virens genomics would lay a solid foundation for the understanding of molecular mechanisms underlying its pathogenicity.Based on the first draft genome of U.virens that has been established previously in our laboratory,a multi-omics study was performed on U.virens across the genome,transcriptome and interactome levels using the bioinformatics approaches in this dissertation.Firstly,the genomes of U.virens and closely related species Metarhizium anisopliae were systematically compared.It was found that the two species have many large regions of synteny and share high protein identities,while the secreted proteins,especially the putative effectors,of U.virens are significantly less conserved as compared with those in M.anisopliae.By analyzing the microsynteny around the effector gene clusters,many effector gene clusters were found in highly diverse genomic islands flanked by conserved regions,implying the specific evolutionary patterns of effector genes.Comparisons between the reference genome and multiple re-sequenced genomes revealed a low level of intraspecific diversity with very low rate of single nucleotide polymorphisms among these different strains.The gene-expression profiles of U.virens during early infection were analyzed using RNA-seq technology.A number of PHI-base gene,secreted protein genes and secondary metabolism genes were identified to be highly induced during the infection process,indicating that they may play important roles in the pathogenicity of U.virens.Secondly,the protein-protein interaction(PPI)network including 20,217 predicted PPIs among 3,305 proteins in U.virens was constructed using the interolog-based and domain-domain interaction(DDI)-based methods.The reliability of the network was assessed by two computational methods based on the GO annotations and expression correlations,and then confirmed by the yeast two-hybrid assays.The predicted network possesses a typical scale-free structure,in which a few hub proteins exist to interact with many proteins.The hub proteins were further divided into two types,static and dynamic hubs,which are functionally and topologically different from each other.A pathogenicity-related subnetwork was also constructed by utilizing the predicted pathogenicity proteins and transcriptome data.Many proteins known to be associated with pathogenicity were found in the subnetwork,while novel pathogenicity proteins could be predicted through data mining.In addition,an inter-species PPI network between U.virens and rice were constructed.It was found that the interacting proteins in U.virens and rice tend to be enriched in different functional categories,and the proteins might exert their functions by targeting the hub proteins in the rice PPI network.At last,U.virens genome was assembled to the chromosome-level using the third-generation sequencing technology.The complete genome is 37.06 Mb in size,and contains seven chromosomes,each containing 2?7 M base pairs.Based on the new version of genome sequence,more repetitive elements were identified as compared with the first version of U.virens genome,and 8,297 protein-coding genes were predicted.Due to the improved assembly quality,the number of genes in important functional categories such as genes encoding secreted proteins and effectors increased,though the total gene number is slightly less than the former one.By comparing the genomes of different U.virens strains,major differences were found to be lied in the regions enriched in repetitive sequences,rather than in the gene-rich regions.Moreover,the complete mitochondrial genomes of several U.virens strains were successfully constructed,which are much longer than those of most phytopathogenic fungi.The intra-species diversity of lengths among different mitochondrial genomes could be explained by the different insertional patterns of homing endonuclease genes.In addition,it is worthy to note that the evolutional relationship based on mitochondrial genomes in different strains was greatly inconsistent with the one according to chromosomal genomes.In summary,this study systematically analyzed U.virens genome sequences,pathogenicity proteins,PPI network,inter-and intra-species differences across several omic layers based on bioinformatics approaches.The findings provide new insights to the pathogenicity mechanisms and evolutionary characteristics in U.virens.