Skip to main content
Fig. 1 | Human Genomics

Fig. 1

From: Network machine learning maps phytochemically rich “Hyperfoods” to fight COVID-19

Fig. 1

Schematic diagram of the overall workflow. The random walk with restarts algorithm operating within a mobile supercomputing DreamLab App is used to simulate how drug and food-based compounds interact with COVID-19-associated viral gene/protein networks. This has been extrapolated from human genome-wide gene-gene (protein-protein) interactome data and based on known COVID-19 human proteome viral targets (i.e. human genes/proteins interacting with different stages of the virus life cycle to facilitate replication and/or enhance viral potency). Both disease and molecular compound impacts are propagated through the interactome network to model the overall cellular response/interactome perturbation. The resulting compound and disease profiles are then correlated to rank compounds according to their network “overlap” with “reference” viral profiles. This approach is based on the assumption that to have an effect, candidate compounds should target the same network component(s) as the one(s) disrupted by the virus. Therapeutic effect can be direct, or indirect, for example where compounds are found to interact with neighbouring network nodes, resulting in subsequent effect propagation to the desired target

Back to article page
\