Discovery Of Disease Comorbidities And Antiviral Drugs Based On Human-viral Protein-protein Interactions

Viral infectious have always been regarded as an crucial public health concern.For example,the coronavirus disease 2019(COVID-19),which took a huge toll on human life and health.The protein-protein interactions(PPIs)between human and viruses play important roles in viral infection and host immune responses.Rapid accumulation of experimentally validated human virus PPIs provides an unprecedented opportunity to investigate the regulatory pattern of viral infection.However,these PPIs are scattered in the published literature and have not been used effectively,thus limiting our understanding of the regulatory patterns of human virus interactions.Objectives: In order to systematically study the laws of human-viral protein interactions,we reviewed the protein interactions reported in the published literature,and systematically characterized the regulatory patterns of viral infection.Interactome analysis was expected to reveal disease comorbidities of viral infection,organizational orientation and potential anti-viral drugs.Methods: Text mining was performed to collect the experimentally validated human virus PPIs.We performed functional enrichment analysis to predict the potential pathways regulated by viral infection based on viral-interacting proteins.In human PPIs,differences in topological features between viral-interacting proteins and other human proteins were compared using Wilcoxon rank-sum tests.Network proximity analyses revealed the diseases comorbidities of viral infection and organizational orientation,further prioritized potential anti-viral drugs.Results: Through text mining and literature review,we collected 27,293 experimentally validated human virus PPIs,covering 8 virus families,140 viral proteins and 6,059 human proteins.Functional enrichment analysis revealed that the proteins interacting with viruses tended to be enriched in cell cycle and immune-related pathways.Moreover,we analyzed the topological features of the viral-interacting proteins and found that they have higher degree,betweenness,and closeness than other human proteins,suggesting which tended to locate in central regions of human PPIs network.Based on network proximity analyses,we revealed the associations between complex diseases and viral infections,and tissue orientation for viral infection.We found that both SRAS-CoV-1 and SARS-CoV-2 were found to be closely related to central nervous system diseases,brain diseases and neurodegenerative diseases.Moreover,we found that SRAS-CoV-1 is significantly associated with adrenal gland,liver and brain,and SARS-CoV-2 may also infect liver,heart and brain.Network analysis also implicated potential antiviral drugs,identifying32 drugs that could potentially be used to develop broad-spectrum antivirals.In addition,we found that ribavirin,mycophenolic acid,mycophenolate mofetil,omacetaxine mepesuccinate and doxorubicin were potential therapeutic drugs for SARS-CoV-1.And,we identified 46 potential therapeutic drugs for SARS-CoV-2,including valrubicin,idarubicin,epirubicin,teniposide,and doxorubicin,and verified them through text mining.Finally,we presented the Human Virus Protein-Protein Interaction database(HVPPI,http://bio-bigdata.hrbmu.edu.cn/HVPPI),that provides experimentally validated human virus PPIs as well as seamlessly integrates online functional analysis tools.Conclusions: Based on the systematic analysis of human-virus PPIs,we performed functional enrichment analysis and network topology features analysis,and identified that viral proteins tend to target cell cycle-and immune-related genes and proteins located in the central regions of the human PPI network,enhancing our understanding of the regulation of human-virus interactions.In addition,based on the network analysis method,we analyzed the organizational orientation,disease comorbidities and potential antiviral drugs of viral infections,which is helpful for the clinical diagnosis of viral infections and antiviral drug development.Finally,to improve the utilization of human-virus PPIs data,we presented the HVPPI database,that provides experimentally validated human-virus PPIs as well as seamlessly integrates online functional analysis tools.In summary,comprehensive understanding the regulatory pattern of human virus interactome will provide novel insights into fundamental infectious mechanism discovery and new anti-viral therapy development.