iRefIndex

iRefIndex integrates protein–protein interaction data from multiple primary databases into a non-redundant index using sequence- and taxonomy-based keys to support comparative and systems-level analysis of interaction networks.

Key Features:

• Unification of Databases: Consolidates interaction data from BIND, BioGRID, DIP, HPRD, IntAct, MINT, MPact, MPPI, and OPHID into a single indexed resource.
• Redundancy Resolution: Generates unique keys for interaction records and participant proteins using primary sequences and taxonomy identifiers combined with the Secure Hash Algorithm (SHA) to group identical records across sources.
• Mapping and Feedback Mechanism: Maps protein database references to current protein sequence records and assigns a mapping score that identifies malformed, deprecated, ambiguous, or unfound references and provides feedback to source databases.
• Independent Retrieval: Produces participant-protein keys that enable retrieval of interaction information independently of the original protein reference identifiers used in source records.
• Flexible Grouping Options: Supports alternative redundant groupings based on gene identifiers or near-sequence identity in addition to exact sequence-based grouping.
• Standard Output Format: Provides the indexed interaction data in PSI-MITAB 2.5 format.

Scientific Applications:

• Interaction dataset curation: Resolves redundant and inconsistent records to improve the quality and consistency of aggregated interaction datasets.
• Comparative interactomics: Enables cross-database comparisons and identification of conserved or database-specific interactions across species using sequence- and taxonomy-based grouping.
• Systems biology: Facilitates construction and analysis of protein interaction networks for systems-level investigations.
• Drug discovery: Supports compilation and evaluation of protein interaction partners relevant to target identification and validation.
• Functional genomics: Aids inference of protein function and functional modules by providing consolidated interaction evidence.

Methodology:

Unique keys for interaction records and participant proteins are generated from primary sequences and taxonomy identifiers using the Secure Hash Algorithm (SHA); protein database references are mapped to current sequence records and assigned a mapping score to flag malformed, deprecated, ambiguous, or unfound references, and identical records are grouped based on sequence and taxonomy information.