WeSA

WeSA scores and ranks bait–prey protein–protein interactions by leveraging large-scale proteomics databases to distinguish true interactions from experimental noise.

Key Features:

• Bait–prey experiment support: Operates on interaction data derived from affinity purification, immunoprecipitation, and proximity labeling experiments.
• Statistical scoring system: Implements a weighted socioaffinity metric to distinguish true protein–protein interactions from false positives.
• Data integration: Processes raw experimental bait–prey data alongside integrated interaction datasets including IntAct, BioGRID, and the BioPlex database to re-order interactions by likelihood.
• Performance metrics: Demonstrates ROC AUC up to 0.93 using CORUM-PDB positives and Negatome negatives, with a best threshold yielding TPR 0.84, FPR 0.11, and Precision 0.98.
• Network-based analysis: Provides network-level analysis and visualization to identify and remove spurious nodes and edges from interaction networks.

Scientific Applications:

• Proteomics quality control: Improves confidence in PPI datasets by distinguishing true interactions from experimental noise.
• Cellular process mapping: Supports mapping of protein interaction networks underlying cellular processes.
• Target identification: Aids identification of potential therapeutic targets by refining interaction evidence.
• Molecular mechanism elucidation and modeling: Facilitates elucidation of disease-related molecular mechanisms and construction of accurate protein network models.

Methodology:

Applies a weighted socioaffinity metric to processed bait–prey experimental data integrated with IntAct, BioGRID, and BioPlex, and evaluates performance using ROC analysis with CORUM-PDB positives and Negatome negatives.