UniProt: a worldwide hub of protein knowledge
By Tunca Dogan. 2018
7,014 citations recorded by the source corpus.
Abstract The UniProt Knowledgebase is a collection of sequences and annotations for over 120 million proteins across all branches of life. Detailed annotations extracted from the literature by expert curators have been collected for over half a million of these proteins. These annotations are supplemented by annotations provided by rule based automated systems, and those imported from other resources. In this article we describe significant updates that we have made over the last 2 years to the resource. We have greatly expanded the number of Reference Proteomes that we provide and in particular we have focussed on improving the number of viral Reference Proteomes. The UniProt website has been augmented with new data visualizations for the subcellular localization of proteins as well as their structure and interactions. UniProt resources are available under a CC-BY (4.0) license via the web at https://www.uniprot.org/.
Subfields: Molecular Biology, Computational Biology & Bioinformatics
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Molecular Biology
- Jinbo Xu
- Burkhard Rost
- David T. Jones
- Yang Zhang
- Peer Bork
- Arne Elofsson
- Liam J. McGuffin
- Paul D. Thomas
In Computational Biology & Bioinformatics
- Burkhard Rost
- Arne Elofsson
- Michael Heinzinger
- Debora Marks
- Martin Weigt
- Lucy J. Colwell
- Jinbo Xu
- Kevin Kaichuang Yang
Citations in the corpus
- Rethinking Attention with Performers (2020, methodology): “A random baseline, with uniform probability across all the vocabulary tokens at every position, has accuracy 5% (when including only the 20 standard amino acids) and 4% (when also including the 5 anomalous amino acids (Consortium, 2019)).”
- GPCRdb in 2021: integrating GPCR sequence, structure and function (2020, methodology): “G protein-coupled receptors (GPCRs) account for 4% (799 (1) / 20 595 (2)) of human genes and mediate the actions of two-thirds (342/515) of hormones and neurotransmitters (3).”
- SPOT-Disorder2: Improved Protein Intrinsic Disorder Prediction by Ensembled Deep Learning (2019, background): “Analysis on the UniProtKB/Swiss-Prot database (as of Dec 2018) [54] has shown that more than 91% of proteins consist of 700 AA residues, indicating that SPOT-Disorder2 is applicable to the vast majority of available sequences.”
- IntFOLD: an integrated web resource for high performance protein structure and function prediction (2019, background): “At the time of writing, we now have close to two hundred mil- lion unique protein sequences in UniProt (1), but the number of protein structures in the Protein Data Bank (PDB) (2) remains 150 000.”
- Evolutionary context-integrated deep sequence modeling for protein engineering (2020, methodology): “However, as these models are trained on massive sequences such as those in UniProt [31] and Pfam [32], the learned representations only capture general context for a wide spectrum of proteins but may not be specific to the protein to be engineered.”
- UDSMProt: universal deep sequence models for protein classification (2019, methodology): “estimate for this number is given by the size of the sparsely annotated TrEMBL dataset (158M) and should be set into perspective by comparison to the size of well-curated SwissProt [1] dataset (560K) with much more complete annotation of protein properties.”
- Evaluating Protein Transfer Learning with TAPE (2019, methodology): “We encode amino acids in the standard 25 -character alphabet, with 20 characters for the standard amino acids, 2 for the non-standard amino acids selenocysteine and pyrrolysine, 2 for ambiguous amino acids, and 1 for when the amino acid is unknown [1, 17].”
- Masked Language Modeling for Proteins via Linearly Scalable Long-Context Transformers (2020, methodology): “For this proof of principle study, we use protein sequences from the Jan. 2019 release of Trembl [14], and concatenated protein sequences to length L = 8192 , long enough to model protein interaction networks without the large sequence alignments required by existing methods [13].”
- Limits and potential of combined folding and docking (2021, methodology): “The default database used is all reference proteomes from UniProt [47] as of May 2020.”
- Protein sequence design with deep generative models (2021, background): “These methods rely on the metagenomic sequencing and subsequent deposition in databases such as UniProt [8], and continued development of these databases are essential for furthering our understanding of biology.”