HAPPI: an online database of comprehensive human annotated and predicted protein interactions

Area of Science:

• Bioinformatics
• Systems Biology
• Molecular Biology

Background:

• Human protein-protein interactions (PPIs) are crucial for understanding molecular signaling and biomolecule functions.
• Existing human PPI databases suffer from limited data coverage and variable quality.
• Integrating PPI data is essential for disease-specific network biology modeling.

Purpose of the Study:

• To develop a novel, web-based resource for human protein-protein interaction data.
• To integrate and standardize information from multiple public PPI databases.
• To provide quality measures for assessing the reliability of protein interactions.

Main Methods:

• Developed the Human Annotated and Predicted Protein Interaction (HAPPI) database.
• Integrated data from HPRD, BIND, MINT, STRING, and OPHID using database integration techniques.
• Employed a unified entity-relationship model and a scoring system for PPI reliability (1-5 star ranks).

Main Results:

• The HAPPI database contains 142,956 non-redundant human PPI pairs among 10,592 proteins (as of summer 2008).
• A subset of HAPPI (star ranks 3-5) demonstrates significantly higher average quality than other human PPI databases.
• The HAPPI web application offers interactive exploration with hyperlinked gene, pathway, and protein domain information.

Conclusions:

• HAPPI represents the most comprehensive public compilation of human protein interaction data.
• The database facilitates exploration of PPI data with quality assessments and annotations.
• HAPPI supports emerging network biology studies in translational bioinformatics.