HVIDB: a comprehensive database for human-virus protein-protein interactions.

Area of Science:

• Virology
• Bioinformatics
• Computational Biology

Background:

• Viral infectious diseases pose a significant public health challenge, necessitating a deeper understanding of host-pathogen interactions.
• Human-virus protein-protein interactions (PPIs) are crucial for viral pathogenesis and represent key targets for antiviral therapies.
• Existing human-virus PPI databases are limited in scope, annotation, and predictive capabilities.

Purpose of the Study:

• To introduce the Human-Virus Interaction DataBase (HVIDB), a comprehensive resource for human-virus PPI data.
• To integrate advanced computational tools for predicting novel human-virus PPIs.
• To facilitate research into viral infection mechanisms and the development of antiviral strategies.

Main Methods:

• Curated and annotated experimentally verified human-virus PPI data.
• Integration of machine learning models for PPI prediction (sequence embedding, interolog mapping, domain-domain interaction inference).
• Inclusion of tissue-specific gene expression profiles and differential gene expression data post-infection.

Main Results:

• HVIDB contains 48,643 experimentally verified human-virus PPIs across 35 virus families.
• The database includes 6,633 targeted host complexes, 3,572 host factors, and 911 3D complex structures.
• Provides tissue-specific expression data for 6,790 human genes and 129 Gene Expression Omnibus series.

Conclusions:

• HVIDB serves as a valuable, one-stop knowledge base for human-virus interactions.
• The integrated prediction tools and rich annotations empower researchers to explore viral pathogenesis.
• This resource is expected to accelerate experimental research and the discovery of novel antiviral therapeutic targets.