UniProtKB: providing protein sequence and functional information to facilitate scientific discovery. 

The data curation of the UniProt Knowledgebase (UniProtKB) is one of main activities of the UniProt Consortium. It involves the evaluation and integration of information from multiple resources so as to represent biological knowledge in the most accurate and comprehensive way. UniProtKB consists of two sections, UniProtKB/Swiss-Prot and UniProtKB/TrEMBL. The former contains manually reviewed records while the latter contains computationally generated records enhanced by automatic classification and annotation. UniProtKB/Swiss-Prot manual curation includes a critical review of experimental and predicted data for each protein. Sequences from the same gene are merged while discrepancies and variants are evaluated case by case. The curation protocol includes manual extraction and structuring of information from the literature, performance and evaluation of computational analyses, and mining and integration of large-scale data sets. Information is updated continuously as new data become available. UniProtKB has also developed UniRule and the Statistical Automatic Annotation System (SAAS), two complementary approaches to automatically annotate protein records in UniProtKB/TrEMBL. Both use InterPro and the manually curated data in UniProtKB/Swiss-Prot as templates to provide information about the many uncharacterised proteins for which there is no experimental data available. Here, we will present the current annotation process in UniProtKB and show how literature-based manual curation coupled with large-scale automatic annotation provides high-quality data across the proteomes provided by UniProtKB, enabling researchers working on comparative and evolutionary genomics to analyse proteomes, and thus genomes, across broad taxonomic ranges. UniProt is updated every four weeks and can be freely accessed or downloaded from https://www.uniprot.org. Programmatic access is also offered through a dedicated API.