A Transomics Study To Discover The Core Molecular Pathogenesis Complex Network System And Construct The Prediction And Diagnosis Model Of Tuberculosis

Objective: The current global prevention and control of tuberculosis(TB)is so severe.And it needs to have a deeper understanding of the pathogenesis and develop more effective prevention strategies and methods.In this study,we hope to find out the pathogenesis complex network system of TB molecular mechanism from the perspective of transomics.Based on this,we could build the prediction and diagnosis model of TB,which could be used in the screening of high risk population of TB and help to achieve precision prevention of TB.At the same time,we hope to demonstrate the relevant evidence of the central dogma in the transomics level,and provide the basis for development of theoretical biology in the future.Methods: This study was a transomics level research based on the big data of multiomics and computer algorithms.Firstly,the data of genome,nucleome,transcriptome and proteome were obtained from international biological database.Secondly,the disease related statistical value of single nucleotide polymorphisms(SNPs)and gene expression were calculated by the conventional analysis process of PLINK and limma respectively.Thirdly,the data of nucleome were integrated into a chromatin-chromatin interaction matrix and a cluster.Then,the TB associated chromatin disease module was obtain by the diseases association analysis combined the statistical value of SNPs and gene expression.Fourthly,the original TB prediction and diagnosis model was constructed by machine learning based on the information of SNPs in the TB associated chromatin disease module.Fifthly,the proteome data were integrated to construct the protein-protein interaction(PPI)network.Using the correlation analysis between matrix of nucleome and proteome,we found the evidence of the central dogma in the transomics level.Finally,we obtained the core and core units of pathogenesis complex network system of TB molecular mechanism by integrated the PPI network and the TB associated chromatin disease module using the network decomposition algorithm.And then,based on these we constructed a new TB prediction and diagnosis model and verified its classification effect with the receiver operating characteristic(ROC)analysis.Results: After the analysis of PLINK based on the genomic data,a total of 49236 SNPs were p<0.05 under the condition of no statistical adjustment.In the transcriptome data analysis,a total of 1594 genes differentially-expressed(DEGs)were identified,of which 738 DEGs were up regulated and 856 DEGs were down regulated.The data of nucleome were integrated into a 3044*3044 unified matrix by separated with 1-22 and X chromatins.The TB associated chromatin disease module was including 101417 SNPs.The area under ROC curve(AUC)of the original TB prediction and diagnosis model was 0.926(sensitivity=0.87,specificity=0.866).We found the evidence that the nucleome is associated with proteome.The pathogenesis complex network system of TB molecular mechanism contains 5846 nodes and 458653 edges,the core of it contains 2015 nodes and 61318 edges that also include 15 core units containing 228 genes.The AUC of new TB prediction and diagnosis model including 2260 SNPs was 0.841(sensitivity=0.768,specificity=0.769).Conclusions: In present study,we initially discovered the core pathogenesis complex network system of TB molecular mechanism by transomics analysis,and constructed the TB prediction and diagnosis model by machine learning that could be a tool for screening high-risk populations of TB.In the aspect of theoretical biology,we have found some clues of the central dogma in the transomics level,which could be a initial exploratory work for the future development of theoretical biology.In terms of the molecular mechanisms of complex diseases,we preliminarily constructed a transomics analysis procedure for general complex diseases and verified that the correlation between genes is important that should be paid attention to in the future research.Finally,because this study was based on the transomics analysis and computer algorithm,the conclusions were needed to be verified by laboratory,clinical and epidemiological researches in future.