The Molecular Interaction Database (MINT) is a public repository for protein-protein interactions (PPI) reported in peer-reviewed journals. As of September 2011, MINT contains approximately 235,000 binary interactions from over 4,750 publications. The database is part of the International Molecular Exchange (IMEx) consortium and uses the PSI-MI standards for curation and data exchange. MINT data are freely accessible and downloadable. MINT has grown significantly over the years, with an average of 4,000 queries per month and thousands of PPI records downloaded annually. The database includes interactions from over 30 species, including humans, mice, and yeast. MINT has implemented a new scoring system to assess the reliability of interactions based on experimental evidence and community recognition. The new system uses a weighted sum of supporting manuscripts to calculate a score, which is mapped to a 0–1 interval. MINT has also introduced a new structural digital abstract (SDA) to facilitate automatic retrieval of protein interaction information. The SDA includes predefined terms and hyperlinks to relevant databases. MINT has also developed a common query interface (PSICQUIC) to standardize programmatic access to molecular interaction databases. MINT's curation policy has evolved to meet the development of protein interaction technology and the evolution of PSI-MI standards. The IMEx consortium has developed a common curation manual specifying which information should be captured and how to represent it. MINT curates articles from FEBS Letters, FEBS Journal, EMBO Journal, and EMBO Reports, and records curated to the IMEx standard are easily recognizable by their IMEx ID. MINT has also introduced a new curation depth and coverage system to classify entries based on the level of detail and the extent of experimental information captured. Entries are labeled with tags indicating the curation depth and coverage. MINT is committed to extending the PSI-MI 2.5 XML format to fully capture the richness of interaction information. MINT is also exploring the use of text mining methods to automatically capture PPI information, as part of the BioCreAtIvE challenge. Results have shown that the coverage and precision of automatic annotation are steadily increasing. MINT is planning to include entries corresponding to another curation level where automatically extracted PPI information is filtered and validated by expert curators before being stored in the database.The Molecular Interaction Database (MINT) is a public repository for protein-protein interactions (PPI) reported in peer-reviewed journals. As of September 2011, MINT contains approximately 235,000 binary interactions from over 4,750 publications. The database is part of the International Molecular Exchange (IMEx) consortium and uses the PSI-MI standards for curation and data exchange. MINT data are freely accessible and downloadable. MINT has grown significantly over the years, with an average of 4,000 queries per month and thousands of PPI records downloaded annually. The database includes interactions from over 30 species, including humans, mice, and yeast. MINT has implemented a new scoring system to assess the reliability of interactions based on experimental evidence and community recognition. The new system uses a weighted sum of supporting manuscripts to calculate a score, which is mapped to a 0–1 interval. MINT has also introduced a new structural digital abstract (SDA) to facilitate automatic retrieval of protein interaction information. The SDA includes predefined terms and hyperlinks to relevant databases. MINT has also developed a common query interface (PSICQUIC) to standardize programmatic access to molecular interaction databases. MINT's curation policy has evolved to meet the development of protein interaction technology and the evolution of PSI-MI standards. The IMEx consortium has developed a common curation manual specifying which information should be captured and how to represent it. MINT curates articles from FEBS Letters, FEBS Journal, EMBO Journal, and EMBO Reports, and records curated to the IMEx standard are easily recognizable by their IMEx ID. MINT has also introduced a new curation depth and coverage system to classify entries based on the level of detail and the extent of experimental information captured. Entries are labeled with tags indicating the curation depth and coverage. MINT is committed to extending the PSI-MI 2.5 XML format to fully capture the richness of interaction information. MINT is also exploring the use of text mining methods to automatically capture PPI information, as part of the BioCreAtIvE challenge. Results have shown that the coverage and precision of automatic annotation are steadily increasing. MINT is planning to include entries corresponding to another curation level where automatically extracted PPI information is filtered and validated by expert curators before being stored in the database.