Multitask learning for host–pathogen protein interactions

Kshirsagar, Meghana; Carbonell, Jaime; Klein-Seetharaman, Judith

The pathway-based objective Biologists often represent the set of human proteins involved in a particular biological process by a graph called a 'biological pathway'. One such example, the 'glucose transport pathway' in human is shown in the Supplementary. To use this pathway construct, we revise our hypothesis to 'proteins from different bacterial species are likely to interact with human proteins from the same biological pathway'. Figure 1B illustrates an example where this hypothesis holds. The pathway information for each human protein can be obtained from pathway databases like Reactome (Matthews et al., 2009) and PID (Schaefer et al., 2009) . While pathways are generally represented as graphs, for our current work we do not use the edges. We treat a pathway as a set of proteins-a human protein h can be a member of several pathways depending on the biological processes it is involved in. Let N be the total number of pathways in human. For a protein pair i ¼ 5b, h4, let p i 2 f0, 1g N be the binary 'pathway vector' indicating the pathway membership of h.