A virus is a much smaller infectious organism than a bacterium. To survive, it must find a suitable host to live inside and use protein-protein interaction (PPI) to dominate the host cell's survival function and absorb nutrients. In turn, host cells use PPI to activate innate antiviral defenses and adaptive immune systems to control viral replication. But in this battle, once the host cell fails to stop the virus from replicating, the large number of "parasitic" viruses that rapidly multiply in its body will tilt out and start a new "parasitic" journey, infecting more healthy cells.
Therefore, knowledge of protein-protein interactions (PPI) is crucial to understanding the relationship between viruses and their hosts. Currently, scientists mainly study protein-protein interactions through high-throughput methods. Although many new discoveries have been made, the lack of scalability of this method also limits further studies. To this end, the researchers developed a new computing framework called P-HIPSTer. It can use the information of protein structure to predict the interaction between virus protein and human protein, which can effectively compensate for the lack of scalability of high-throughput method.
The researchers say p-hipster has been used to study more than 2,000 known viruses that infect humans and about 23,000 proteins that encode them. The algorithm predicted about 385,000 possible interacting protein pairs with an accuracy of nearly 83 percent, creating a small database of human-virus protein interactions. It has contributed to the further research of human immunology and infectious diseases.
Dr. Sagi Shapira, a professor in the department of microbiology and immunology at Columbia University, who led the study, said the database of human-viral protein interactions created by the institute could help scientists break through current research bottlenecks to better understand viral and human protein interactions. Next, P-HIPSTer will be used to study more complex pathogens, such as parasites and bacteria.