The PROSITE database is a collection of biologically meaningful signatures, including patterns and profiles, used to identify conserved regions in proteins. Each signature is linked to documentation providing information on the protein family, domain, or functional site. The database has been updated with new rules to improve precision. The latest version (release 19.11) contains 1329 patterns and 552 profile entries. Over two years, more than 200 domains have been added, and 52% of UniProtKB/Swiss-Prot entries now reference PROSITE. The database is accessible at http://www.expasy.org/prosite/. PROSITE uses two types of signatures: patterns and profiles. Patterns are useful for identifying short, well-conserved regions but require regular updates to account for new variability. A tool has been developed to automatically update patterns by minimizing false positives (FP) and false negatives (FN). This process involves selecting patterns that can be updated, detecting mismatches, and refining patterns to ensure they remain stringent. The updated patterns are then tested and manually checked for FP. This strategy has improved the accuracy of the database, reducing FP and FN by about 25% in Swiss-Prot entries. ProRule is a new database that stores functional information from PROSITE in a format that can be used by programs. It provides general and conditional information for annotation. General information includes domain names and boundaries, while conditional information depends on specific amino acids or domain arrangements. ProRule is used by Swiss-Prot curators and is accessible to external users. The PROSITE website has been redesigned to include new predictive tools for detailed functional information. Matches on UniProtKB and PDB entries are pre-calculated and stored in a relational database. A domain visualizer is integrated into the results, allowing the identification of functional residues. The website also provides access to ProRule and other tools. The profile method construction has been improved to handle more divergent proteins. New strategies allow for more specific annotations based on residue-level information. Profile builder parameters are adjusted according to annotations, improving the accuracy of local alignments. PROSITE and ProRule are freely available to academic users. Commercial users must contact the Swiss Institute of Bioinformatics for a license. PROSITE is updated weekly on an FTP server and is accessible via various web pages. The database is supported by the Swiss National Science Foundation and funded by the State Secretariat for Education and Research.The PROSITE database is a collection of biologically meaningful signatures, including patterns and profiles, used to identify conserved regions in proteins. Each signature is linked to documentation providing information on the protein family, domain, or functional site. The database has been updated with new rules to improve precision. The latest version (release 19.11) contains 1329 patterns and 552 profile entries. Over two years, more than 200 domains have been added, and 52% of UniProtKB/Swiss-Prot entries now reference PROSITE. The database is accessible at http://www.expasy.org/prosite/. PROSITE uses two types of signatures: patterns and profiles. Patterns are useful for identifying short, well-conserved regions but require regular updates to account for new variability. A tool has been developed to automatically update patterns by minimizing false positives (FP) and false negatives (FN). This process involves selecting patterns that can be updated, detecting mismatches, and refining patterns to ensure they remain stringent. The updated patterns are then tested and manually checked for FP. This strategy has improved the accuracy of the database, reducing FP and FN by about 25% in Swiss-Prot entries. ProRule is a new database that stores functional information from PROSITE in a format that can be used by programs. It provides general and conditional information for annotation. General information includes domain names and boundaries, while conditional information depends on specific amino acids or domain arrangements. ProRule is used by Swiss-Prot curators and is accessible to external users. The PROSITE website has been redesigned to include new predictive tools for detailed functional information. Matches on UniProtKB and PDB entries are pre-calculated and stored in a relational database. A domain visualizer is integrated into the results, allowing the identification of functional residues. The website also provides access to ProRule and other tools. The profile method construction has been improved to handle more divergent proteins. New strategies allow for more specific annotations based on residue-level information. Profile builder parameters are adjusted according to annotations, improving the accuracy of local alignments. PROSITE and ProRule are freely available to academic users. Commercial users must contact the Swiss Institute of Bioinformatics for a license. PROSITE is updated weekly on an FTP server and is accessible via various web pages. The database is supported by the Swiss National Science Foundation and funded by the State Secretariat for Education and Research.