Tuberculosis-Related Gene Prediction And Network Analysis Based On Multifaceted Information Fusion

Tuberculosis(TB)is one of the biggest infectious disease killers caused by Mycobacterium tuberculosis(MTB).Although intensive efforts have been devoted to exploring the pathogenesis of this disease,its systematic understanding of the corresponding host immune response remains poor understood.The main reason is that the related target genes have not been completely excavated,so the underlying molecular mechanism of MTB interfering with its host cell cannot be systematically outlined.Studying the protein-protein interactions(PPIs)between MTB and human could establish linkages between MTB and provide informative candidate target genes,which are valuable for discovery of the response mechanism of human against MTB infection.However,the known MTB-human protein interactions are very scarce.The primary reason might be that the experimental determination of PHIs is time-consuming and labor-intensive.Therefore,it is highly needed to develop computational methods to guide or aid experimental techniques for identifying MTB-human PPIs.Additionally,it is estimated that more than 2 billion individuals are infected with MTB worldwide,among which Approximately 85-90% of infected individuals are latent tuberculosis(LTB)infection.Among patients with LTB,approximately 10% will progress to active pulmonary TB(PTB).Although intensive efforts have been devoted to investigating latent tuberculosis(LTB)and active tuberculosis(PTB)infections,the similarities and differences in the host responses to these two closely associated stages remain elusive.In particular,it would be difficult to decode the systemic response induced by LTB from an experimental perspective.One of the major obstacles is that most host genes show trivial expression changes in the LTB stage,which makes it difficult to identify informative genes correlated with LTB using traditional methods such as differential gene expression analysis.Therefore,a novel research strategy or framework,especially from a computational viewpoint,is urgently needed to compare PTB and LTB infections.To address these issues,the first part in this study we proposed an integrated framework that combined template-,domain-domain interaction-,and machine learning-based methods to predict MTB-human protein-protein interactions.As a result,we established a network composed of 13,758 PPIs including 451 MTB proteins and 3,167 human proteins.We demonstrate that the integrated strategy can effectively reduce the false positive predictions.Extensive validations based on different resources verify that the predicted MTB-human PPI network is largely reliable.Compared to known human targets of various pathogens,our predicted human targets show a similar tendency in terms of the network topological properties and enrichment in important functional genes.Additionally,these human targets largely have longer sequence lengths,more protein domains,more disordered residues,lower evolutionary rates,and older protein ages.Functional analysis demonstrates that the human target proteins are mainly related to phosphorylation,kinase activity,and signaling transduction,while the pathway analysis demonstrates that these proteins are highly involved in the disease and immune related pathways.Dissecting the cross-talk among top-ranked pathways suggests that the cancer pathway may serve as a bridge in MTB infection.Triplet analysis illustrates that the paired targets interacting with the same partner are adjacent to each other in the intra-species network and tend to share similar expression patterns.Finally,we identified 36 potential anti-MTB human targets by integrating known drug target information and molecular properties of proteins.The second part in this paper we developed a framework known as the consistently differential expression network(CDEN)to identify TB-related gene pairs by combining microarray profiles and protein-protein interactions.The differential interaction-based approaches not only integrated transcriptional information with molecular networks but also considered the discrepancies in the expression correlations of gene pairs rather than the expression changes in genes between different disease states.This strategy makes it possible to uncover the immune responses to the LTB stage,which would further facilitate the study of the similarities and specific differences in the responses to LTB and PTB infections.We thus obtained 774 and 693 pairs corresponding to the PTB and LTB stages,respectively.Furthermore,we systematically compared the CDENs of LTB and PTB at the node,edge,and module levels.By investigating the molecular properties,we found that the common genes observed in these two stages were more evolutionarily conserved,topologically important,and posttranslationally modified than the PTB-and LTB-specific genes,and trivial differences between the two categories of specific genes were demonstrated.Regarding the gene expression measurements,PTB-specific genes showed higher expression levels than their LTB-specific counterparts.Moreover,PTB-specific genes revealed the most significant fold-changes among the three categories of PTB-LTB genes in the PTB stage,implying that the fluctuations in the expression of these genes are tightly correlated with the progression of TB.The distributions of the expression correlations of the detected gene pairs revealed that LTB-related interactions generally had lower PCCs and would be more suppressed,whereas PTB-related interactions generally had higher PCCs and would be more activated.The MIR analyses demonstrated that our gene pairs preferred to cluster within the topological and functional modules.Based on the pathway enrichment analysis,the genes in the CDENs were highly involved in pathogen infections and immune responses.Additionally,choosing the NF-?B signaling pathway as a representative,we observed that the PTB-specific genes tended to be in upstream positions within this pathway,whereas the LTB-specific genes tended to appear in downstream positions.Finally,we identified existing drugs that were associated with the gene pairs and found several pairs in which both genes were the successful targets of existing drugs,which provided novel insights into the specific treatment of TB patients in different stages.