PNNS

PNNS computes a PageRank-based affinity to identify proteins functionally closest to a queried node within protein-protein interaction (PPI) networks.

Key Features:

• PageRank Affinity Measure: Computes a PageRank Affinity derived from personalized PageRank, quantifying closeness by measuring the frequency a smaller-degree protein is visited during random walks that restart at a larger-degree protein.
• Robustness to Noise: Considers paths of all lengths via personalized PageRank, providing resilience to noisy or incomplete PPI data.
• Biological Significance: The PageRank Affinity correlates with cluster co-membership and serves as a predictor of co-complex membership and functionally related proteins.
• Scalability: Implements an algorithm that scales to large biological networks, enabling analysis of extensive PPI datasets such as BioGRID or user-specified networks.

Scientific Applications:

• Proteome functional organization: Identifies proteins closely related in function to a query protein within PPI networks to explore proteome organization.
• Co-complex membership prediction: Predicts membership of proteins in the same complex based on PageRank Affinity.
• Identification of functionally related proteins: Detects proteins with similar functional roles by ranking nodes according to PageRank-based closeness.

Methodology:

Applies personalized PageRank (random walks with restarts) on PPI networks, considering paths of all lengths to compute the PageRank Affinity measure.