The AAindex database is a collection of numerical indices representing physicochemical and biochemical properties of amino acids and amino acid pairs. In 2008, the database was updated to include a new section, AAindex3, which contains statistical protein contact potentials. The database now consists of three sections: AAindex1 for amino acid indices, AAindex2 for amino acid substitution matrices, and AAindex3 for protein contact potentials. All data are derived from published literature. The database can be accessed through the DBGET/LinkDB system at GenomeNet or downloaded via anonymous FTP. AAindex has been used in various bioinformatics studies, including predicting protein subcellular localization, immunogenicity of MHC class I binding peptides, and coordinated substitutions in protein sequences. A derivative database, UMBC AAindex Database, and a web tool for visualizing relationships among AAindex entries are also available. In 2005, Pokarowski et al. compared 29 published matrices of protein pairwise contact potentials. These potentials have been used to predict protein structures. They found that each contact potential is similar to one of two popular matrices derived by Miyazawa and Jernigan. Recently, the same team found intermediate links between substitution matrices and contact potentials, with mutual correlations of at least 0.8. These relations provide motivation to extend the AAindex database. The AAindex database is released annually, with the latest version being 9.0. AAindex1 contains 544 amino acid indices, each with an accession number, description, reference information, and numerical values for the properties of 20 amino acids. AAindex2 contains 94 amino acid substitution matrices, including symmetric and nonsymmetric matrices. AAindex3 contains 47 amino acid contact potential matrices, including symmetric and nonsymmetric matrices. The database is available through the DBGET/LinkDB system or anonymous FTP. It is also supported by BioRuby and EMBOSS. Users are requested to cite this article when using the AAindex database.The AAindex database is a collection of numerical indices representing physicochemical and biochemical properties of amino acids and amino acid pairs. In 2008, the database was updated to include a new section, AAindex3, which contains statistical protein contact potentials. The database now consists of three sections: AAindex1 for amino acid indices, AAindex2 for amino acid substitution matrices, and AAindex3 for protein contact potentials. All data are derived from published literature. The database can be accessed through the DBGET/LinkDB system at GenomeNet or downloaded via anonymous FTP. AAindex has been used in various bioinformatics studies, including predicting protein subcellular localization, immunogenicity of MHC class I binding peptides, and coordinated substitutions in protein sequences. A derivative database, UMBC AAindex Database, and a web tool for visualizing relationships among AAindex entries are also available. In 2005, Pokarowski et al. compared 29 published matrices of protein pairwise contact potentials. These potentials have been used to predict protein structures. They found that each contact potential is similar to one of two popular matrices derived by Miyazawa and Jernigan. Recently, the same team found intermediate links between substitution matrices and contact potentials, with mutual correlations of at least 0.8. These relations provide motivation to extend the AAindex database. The AAindex database is released annually, with the latest version being 9.0. AAindex1 contains 544 amino acid indices, each with an accession number, description, reference information, and numerical values for the properties of 20 amino acids. AAindex2 contains 94 amino acid substitution matrices, including symmetric and nonsymmetric matrices. AAindex3 contains 47 amino acid contact potential matrices, including symmetric and nonsymmetric matrices. The database is available through the DBGET/LinkDB system or anonymous FTP. It is also supported by BioRuby and EMBOSS. Users are requested to cite this article when using the AAindex database.