DOMINE

DOMINE compiles known and predicted domain-domain interactions (DDIs) to support analysis of protein interaction networks and inference of protein-protein interactions.

Key Features:

• Data integration: Integrates DDI data from 15 diverse sources.
• Structure-derived inferences: Adds 2,285 DDIs inferred from experimentally characterized high-resolution three-dimensional protein structures.
• Computational predictions: Incorporates approximately 3,500 novel predictions produced by five distinct computational methodologies developed over the past three years.
• Coverage: Contains 26,219 unique DDIs among 5,140 unique Pfam domains, representing a 23% increase from the previous 20,513 DDIs across 4,346 domains.
• Confidence classification: Implements a novel classification scheme that assigns confidence levels to predicted DDIs.

Scientific Applications:

• Experimental validation: Supports bench scientists in testing and validating proposed domain and protein interactions.
• PPI prediction: Enables bioinformaticians to predict novel protein-protein interactions based on existing DDIs.

Methodology:

Integration of 15 data sources; inference of 2,285 DDIs from experimentally characterized high-resolution 3D protein structures; generation of ~3,500 predictions by five computational methodologies; and application of a classification scheme to assign confidence levels to predicted DDIs.