The UniProt Knowledgebase (UniProtKB) aims to provide the scientific community with a comprehensive, consistent and authoritative resource for protein sequence and functional information. Given the importance of human and vertebrate model data in biomedical research, a major focus is the high-quality manual curation of human proteins and their vertebrate orthologues. Manual curation involves (1) the extraction of experimental results from scientific literature to enrich protein records with a wide range of information including function, structure, interactions and subcellular location, (2) the manual verification of each sequence and clarification of discrepancies between sequence reports, and (3) the assessment of the output of a range of analysis programmes to ensure that sequence features are correctly reported. Manual curation also facilitates the standardization of experimental data – a step necessary for development of methods that enable the semi-automated transfer of manual annotation to uncharacterised or related proteins. Consequently, manual curation of vertebrate proteins plays a vital role in providing users with a complete overview of available data while ensuring its accuracy, reliability and accessibility. UniProtKB/Swiss-Prot currently contains the complete manually reviewed human proteome, comprising approximately 20’300 proteins, and an additional 61’000 reviewed entries from model vertebrates such as mouse, rat, apes, cow, chicken, zebrafish and Xenopus. Ongoing efforts continue to improve the quality of vertebrate sequences in collaboration with HAVANA, Ensembl, HGNC and RefSeq, to include new functional information as it becomes available, and to extend the coverage of curated proteins in vertebrate species. All data are freely available from "http://www.uniprot.org":www.uniprot.org.