MobiDB

MobiDB annotates intrinsic protein disorder and mobility by aggregating experimental evidence, structure-derived inferences, and computational predictions across UniProt and UniParc sequences (over 80 million entries) to support analysis of intrinsically disordered proteins.

Key Features:

• Annotation Levels: Three annotation levels—manually curated from DisProt, indirectly inferred from PDB structures, and predicted using a suite of ten algorithms (including ESpritz, IUPred, DisEMBL, GlobPlot, VSL2b, and JRONN).
• Consensus Annotation: Consensus annotations for long disordered regions integrate multiple sources to classify regions into flexible or constrained disorder categories.
• Integration with External Databases: Annotations incorporate UniProt (post-translational modifications and linear motifs), Pfam (protein family annotations), and STRING (experimental protein–protein interactions classified by disorder content).
• Advanced Search and Customization: Support for constructing custom datasets and exporting annotations in TSV, Fasta, or JSON formats for downstream analyses.
• Programmatic Access (API): An API provides programmatic access to MobiDB data and annotations.
• Novel Annotations: Includes annotations of binding modes of disordered proteins, disorder-to-order transitions, liquid–liquid phase separation, and post-translational modifications.
• Comprehensive Coverage: Covers the complete UniProt protein set and all UniParc sequences, totaling over 80 million entries.
• Statistical Insights: Provides statistics and summary information at both database and proteome levels.

Scientific Applications:

• Intrinsic Disorder Characterization: Analysis of intrinsically disordered proteins (IDPs) and disordered regions across proteomes.
• Protein–Protein Interaction Analysis: Study of interactions and interaction propensity using disorder-classified experimental interactions from STRING.
• Post-Translational Modification and Motif Studies: Investigation of PTMs and linear motifs mapped from UniProt in the context of disorder.
• Conformational Dynamics and Binding Mode Research: Examination of disorder-to-order transitions, binding modes, and liquid–liquid phase separation phenomena.
• Computational Dataset Generation: Creation of custom annotation datasets for downstream computational analyses and modelling.

Methodology:

Annotations combine manual curation from DisProt, structure-derived inferences from PDB entries, and predictions from ten disorder predictors (including ESpritz, IUPred, DisEMBL, GlobPlot, VSL2b, and JRONN); consensus annotation integrates sources to classify long disordered regions as flexible or constrained; annotations are integrated with UniProt, Pfam, and STRING and summarized with database- and proteome-level statistics.