Project description
Applying deep learning to investigate the coronavirus–host interaction
The COVID-19 pandemic has taught us that it is necessary to understand the virus's interaction with the host in order to design effective therapeutics. The EU-funded RiPCoN project will analyse protein–protein interactions and protein–RNA interaction predictions between virus and host and feed this information into an existing deep learning model. This will generate a public resource for translational and basic coronavirus research and help identify approved drugs that are likely effective against 2019-nCoV. Most importantly, scientists will examine how genetic variations in both humans and the virus are jointly responsible for disease severity, hoping to improve risk management and preparedness for future outbreaks.
Objective
We aim to identify approved drugs that can be repurposed for the treatment of 2019-nCoV using interactome profiling and deep-learning. We will deploy rapid high-throughput protein-protein interaction mapping and computational protein-RNA interaction predictions to chart the coronavirus host interactome network (CoHIN), which will become a public resource for translational and basic coronavirus research few months after project start. CoHIN will serve as input into an existing deep-learning model to identify approved drugs that are likely effective against 2019-nCoV, which will be validated in in vitro and in vivo systems. In the second stage we will experimentally determine the matrix of viral protein alleles vs. variants of the interacting human proteins to understand how human and viral natural variations jointly mediate disease severity in different individuals. These data will be integrated with epidemiological and human genomics data to improve risk management and improve preparedness for future coronavirus outbreaks. Overall, we aim to achieve the following objectives: - Map the protein interactome of 2019-nCoV and related Coronaviridae with their human host - Generate the allele interaction matrix and relate differences to epidemiological data - Develop a microarray-based patient screen to detect exposure to 2019-nCoV and identify immunogenic epitopes - Identify 10 approved drugs that are most likely efficient against 2019-nCoV using network integration and deep-learning - Validate drug candidates in in vitro and in vivo systems
Fields of science
Keywords
Programme(s)
Funding Scheme
RIA - Research and Innovation actionCoordinator
85764 Neuherberg
Germany