The Conserved Domain Database (CDD), maintained by the National Center for Biotechnology Information (NCBI), is a public resource for protein annotation based on conserved domain footprints. It has entered its 15th year and continues to improve coverage and consistency in domain annotation. CDD provides both live search functionality and pre-computed domain annotations for sequences in the NCBI Entrez protein database. It integrates domain models and definitions from public and in-house curation efforts, aiming to enhance coverage and provide finer-grained classifications of protein domains. CDD curators generate alignment models that align with domain boundaries observed in 3D protein structures and model structurally conserved cores of domain families. CDD is accessible at http://www.ncbi.nlm.nih.gov/Structure/cdd/cdd.shtml. CDD includes 46,675 protein and domain models from various sources, including Pfam, SMART, COGs, TIGRFAMs, and NCBI's in-house curation. Recent updates include the addition of new domain families and improved clustering methods. CDD is part of NCBI's Entrez system and is cross-linked with other databases. It annotates over 120 million sequences in Entrez/protein, with 98% of structure-derived sequences over 30 residues long. CDD curators manually review domain clusters to ensure accuracy and avoid false clustering. CDD also improves domain annotation by considering domain architecture contexts. A new option in CD-Search allows for higher sensitivity and post-processing of results, enabling the detection of domain hits that may be borderline significant. Structural motifs, such as beta-propellers and coiled coils, are now recorded and mapped, providing more detailed annotations. CDD curators also record functional motifs, including active sites and binding sites, to facilitate the interpretation of sequence conservation and variation. CDD is continuously updated with new domain models and annotations, and it is used for both research and educational purposes. The database is supported by the National Institutes of Health and is freely accessible.The Conserved Domain Database (CDD), maintained by the National Center for Biotechnology Information (NCBI), is a public resource for protein annotation based on conserved domain footprints. It has entered its 15th year and continues to improve coverage and consistency in domain annotation. CDD provides both live search functionality and pre-computed domain annotations for sequences in the NCBI Entrez protein database. It integrates domain models and definitions from public and in-house curation efforts, aiming to enhance coverage and provide finer-grained classifications of protein domains. CDD curators generate alignment models that align with domain boundaries observed in 3D protein structures and model structurally conserved cores of domain families. CDD is accessible at http://www.ncbi.nlm.nih.gov/Structure/cdd/cdd.shtml. CDD includes 46,675 protein and domain models from various sources, including Pfam, SMART, COGs, TIGRFAMs, and NCBI's in-house curation. Recent updates include the addition of new domain families and improved clustering methods. CDD is part of NCBI's Entrez system and is cross-linked with other databases. It annotates over 120 million sequences in Entrez/protein, with 98% of structure-derived sequences over 30 residues long. CDD curators manually review domain clusters to ensure accuracy and avoid false clustering. CDD also improves domain annotation by considering domain architecture contexts. A new option in CD-Search allows for higher sensitivity and post-processing of results, enabling the detection of domain hits that may be borderline significant. Structural motifs, such as beta-propellers and coiled coils, are now recorded and mapped, providing more detailed annotations. CDD curators also record functional motifs, including active sites and binding sites, to facilitate the interpretation of sequence conservation and variation. CDD is continuously updated with new domain models and annotations, and it is used for both research and educational purposes. The database is supported by the National Institutes of Health and is freely accessible.