Protein interactions between a pathogen and its host are fundamental in the establishment of the pathogen and underline the
infection mechanism. In the present work, we developed a single predictive model for building a host-viral interactome based on the identification of structural descriptors from motif-domain interactions of
protein complexes deposited in the
Protein Data Bank (PDB). The structural descriptors were used for searching, in a database of
protein sequences of human and five clinically important viruses; therefore, viral and human
proteins sharing a descriptor were predicted as interacting
proteins. The analysis of the host-viral interactome allowed to identify a set of new interactions that further explain molecular mechanism associated with
viral infections and showed that it was able to capture human
proteins already associated to
viral infections (human infectome) and
non-infectious diseases (human diseasome). The analysis of human
proteins targeted by
viral proteins in the context of a human interactome showed that their neighbors are enriched in
proteins reported with differential expression under
infection and disease conditions. It is expected that the findings of this work will contribute to the development of systems biology for
infectious diseases, and help guide the rational identification and prioritization of novel drug targets.