Analysis Of MiRNAs Mediated Virus-host Network

MiRNAs are a class of endogenous single strand RNAs, which have approximately 22 nucleotides. MiRNAs can bind to 3'untranslated region of transcripts and then cause target degradation or translational block. There are more and more findings showing that miRNAs play an indispensable role in human virus interactions. Either virus or human miRNAs can target transcripts of both human and virus. Viral miRNAs can mediate the network regulated by host miRNAs, and the mechanism of host defenses for virus related RNAi and IFN. A clear understanding of miRNA mediated cross-talks between host and viruses will be important for the thoroughly understanding of the mechanism of host and virus interactions and the development of antiviral drugs.In this thesis, we described the virus-host interaction in the level of miRNAs in two aspects: (1)The regulation principles of herpesvirus miRNAs based on the human protein interaction network (PIN); (2)The co-regulation principles of viral proteins and human miRNAs.First of all, herpesvirus miRNAs regulation principles were analyzed in the framework of human PIN. The ever-increasing data of known viral miRNAs and human PIN have made it possible to study the characteristics of viral miRNA targets in the context of these networks. Here, only viruses belong to Herpesviridae were chosen to do the analyses for the most fully identification of miRNAs of that viruses. Then, human PIN was constructed and all miRNAs belong to six herpesviruses were used to predict the targets on human genes. The analyses were made in the following aspects: (1) Firstly, we inspected whether the targets of herpesvirus miRNAs cover more hubs than random conditions, the results demonstrated that herpesvirus miRNAs preferentially target hub nodes. Secondly, the nodes targeted by herpesvirus miRNAs were classified into two groups: hubs (bottlenecks) and non-hubs (non-bottlenecks). The regulation strength of the herpesvirus miRNAs in the two groups was examined. Statistical significance of virus miRNAs and number of virus types for the two groups were studied, with positive results for both. As hubs are the crucial nodes in the PIN, preferential hub targeting may make herpesvirus miRNAs more efficient in the context of that network.(2) We defined nodes targeted by all six viruses as common targets and nodes targeted by only one virus as specific targets. There were significant differences in degree and betweenness centrality between those common and specific targets. Specifically we observed a significant positive correlation between virus types and the proportion of hubs. It is consistent with the observation that common targets have a higher degree and betweenness centrality than those of specific targets. We propose that common targets might be related to the common pathogenesis processes for all viruses and that specific targets are involved in the specific infection processes for a particular type of viruses. (3) We used k-core and excess retention (ER) analysis to measure the distance between common targets and the topological center. With an increasing k value, higher ER values were indicative of common targets being closer to the global center of the PIN. We defined two parameters for measuring the significance of the module characteristics of common targets, one is the number of nodes in the giant connected component (GCC) comprised of common targets, the other is the density of the sub-network formed by those targets. Compared with random conditions, the GCC and the density of the sub-network formed by common targets have significantly higher values, indicating the module characteristic of these targets. This strategy would facilitate rapid transmission of information to the rest of the network, thereby maximizing viral control of cellular functions. Overall, These results will help unravel the complex mechanism of herpesvirus-host interactions and may provide insight into the development of novel anti-herpesvirus drugs.At the same time, the co-regulation principles of virus and host miRNAs were analyzed. The basic analysis of virus-host protein-protein interactions showed that researches are currently focused on a few viruses. The top ten viral host protein interactions account for 78.59% of the total. Most of host proteins interact with a small number of virus types, only a small number of host proteins interact with a large number of viruses, suggesting that different viruses may have different pathogenesis mechanisms, in spite of a small amount of overlaps among viruses. The construction and analysis of virus relationship network (VRN) and miRNA relationship network (MRN) show that the two networks display scale-free characteristic. The construction of VRN makes the establishment the viral relationships available from the view of PPI network and viruses which has a similar pathogenesis showed closer network distance. Moreover, combined with known viral information and network topology, the characteristic of unknown virus may be predicted. In addition, large GCC can be formed by VRN and MRN, with the fact that the number of viral and miRNAs targets are small. This indicates that large GCC may be archived by the relationship among viruses with similar pathogenesis. To explore the relationship between viral factors and host miRNAs in systems level, the overlapping targets between viral proteins and host miRNAs were analyzed. The results show that the overlap was significant, which hints viruses may evolve to target proteins which are also targeted by host miRNAs. The regulatory mode between viral proteins and host miRNAs in the framework of PIN was next examined and found that viral proteins and host miRNAs tend to target protein pairs interacting with each other. Considering that two target sets overlapped significantly, suggesting that they have a closer distance on PIN, this may be a strategy for viruses to manipulate host or a mechanism for host anti-virus defense, and these maneuvers may be formed in the long-term co-evolution. The topological analysis of the overlapping targets between viral protein and host miRNAs show that these targets tend to be hubs or bottlenecks. K-core analysis revealed overlapping proteins located in the global center of the PIN. These results suggest that viral protein and miRNAs may tend to target key genes in the center network to effectively regulate the cell. Finally, we consider the topological information and GO of both viral protein and host miRNAs targets to predict the potential functional virus-host miRNA pairs. Prediction results may provide reference for subsequent experiments. Finally, a brief summary was made and further possible researches were discussed.