LPCF

LPCF identifies local functional communities in protein-protein interaction (PPI) networks to reveal functional modules associated with queried proteins.

Key Features:

• Local Community Detection: Focuses on local community detection around a queried protein instead of globally partitioning the entire network.
• Algorithms (Nibble and PageRank-Nibble): Implements Nibble and PageRank-Nibble local clustering algorithms that initiate short random walks from the queried protein vertex and identify communities based on popular destinations of these walks.
• Quality Assessment via Conductance: Assesses cluster quality using conductance, described as the proportion of outgoing edges relative to incoming ones, and applicable in undersampled networks.
• Biological Coherence: Targets communities that are functionally coherent and correspond to genuine biological modules.
• Data Sources: Leverages BioGRID and user-specified PPI networks as input.
• Performance Comparison: Has been benchmarked against other graph partitioning algorithms, demonstrating superior performance in identifying high-quality communities.

Scientific Applications:

• Functional Module Investigation: Explores functional modules associated with specific proteins to inform their roles in cellular processes.
• Biological Insight Generation: Identifies previously unknown functional ties between proteins to aid discovery of novel pathways and mechanisms.
• Network Analysis: Analyzes PPI networks from various organisms to support comparative studies and evolutionary biology research.

Methodology:

Implements Nibble and PageRank-Nibble local clustering algorithms that perform short random walks from a queried protein vertex, identify communities based on frequent walk destinations, and evaluate clusters using conductance; performance has been benchmarked against other graph partitioning algorithms.