BIPS

BIPS predicts protein-protein interactions (PPIs) by leveraging interolog information to infer potential intra- and interspecies interactions from known interactions of orthologous genes.

Key Features:

• Interolog-based inference: Infers PPIs by transferring interaction evidence from orthologous genes across species.
• BIANA integration: Consolidates data from multiple PPI-related databases within the BIANA framework to enrich prediction datasets.
• Input flexibility and partner restriction: Accepts sequences or protein identifiers and allows restricting predicted partners to a user-defined list or to the entire BIANA database.
• Metadata integration: Incorporates additional metadata to refine the reliability of predicted interactions.
• Performance evaluation: Assessed using known PPIs from various interactomes with a leave-one-out approach, reporting specificity between 72% and 98% and sensitivity between 1% and 59% depending on sequence identity cut-off thresholds.

Scientific Applications:

• Interactome mapping: Supports reconstruction and extension of intra- and interspecies interactomes using inferred interactions.
• Systems biology: Enables analysis of biological networks by providing candidate PPIs for network modeling.
• Functional genomics: Assists in assigning or predicting protein functions via inferred interaction partners.
• Drug discovery: Provides candidate PPIs that can inform target identification and mechanistic hypotheses.

Methodology:

Interolog transfer using orthology and sequence similarity with sequence identity cut-off thresholds; integration of multiple PPI databases via the BIANA framework; optional restriction of predicted partners to user-defined lists or the full BIANA database; incorporation of additional metadata to refine predictions; evaluation using a leave-one-out approach on known interactomes.