LocDB

LocDB provides experimentally annotated subcellular localizations, expressed in Gene Ontology (GO) terms, for proteins from Homo sapiens (HS) and Arabidopsis thaliana (AT) to support interpretation of protein function and development of localization prediction methods.

Key Features:

• Coverage: Experimental annotations for 19,604 UniProt proteins (13,342 HS; 6,262 AT).
• Ontology annotation: Localizations are expressed using Gene Ontology (GO) terminology.
• Integration of predictions: Includes localization predictions generated by state-of-the-art computational methods.
• Experimental evidence: The type of experimental information supporting each annotation is provided when available.
• Novel annotations: Contributes approximately 10,000 new experimental localization annotations for human proteins and ~900 for Arabidopsis thaliana compared with other public databases.
• Multiplicity: Over 40% of proteins have multiple localization annotations.
• Identifier mapping: Entries are associated with UniProt identifiers and include mappings to Ensembl and TAIR where applicable.

Scientific Applications:

• Functional interpretation: Use experimental localization data to infer protein function within cellular compartments.
• Training and benchmarking: Provide GO-encoded experimental data and evidence types for training and benchmarking computational localization predictors.
• Model coverage improvement: Support development of methods that increase coverage and accuracy of protein localization predictions, aided by many multi-localized proteins.
• Comparative analysis: Enable cross-species comparisons of localization patterns between Homo sapiens and Arabidopsis thaliana.

Methodology:

Annotations are encoded in Gene Ontology (GO) terms and the database integrates localization predictions generated by state-of-the-art computational methods.