AdaPPI

AdaPPI identifies novel protein functional modules within protein-protein interaction (PPI) networks to reveal protein complexes and biological pathways.

Key Features:

• Adaptive Convolution Graph Network: Employs an adaptive convolution graph network tailored to PPI networks to predict protein functional modules by modeling complex nonlinear relationships and handling unevenly sparse and dense node distributions.
• Attributed Graph Node Presentation Algorithm: Integrates protein gene ontology attributes with network topology to adaptively aggregate structural information from low- and high-order graphs based on node distribution while considering graph node smoothness.
• Core Cliques and Expansion Algorithms: Applies core cliques and expansion algorithms to node representations to identify functional modules, including protein complexes and biological pathways.
• Performance Evaluation: Comprehensive performance evaluations and case studies report that AdaPPI outperforms existing state-of-the-art methods in predicting protein functional modules.

Scientific Applications:

• Systems biology: Identifying protein functional modules in PPI networks to elucidate cellular organization and processes.
• Disease mechanism analysis: Characterizing altered protein complexes and pathways to support studies of disease mechanisms and potential therapeutic targets.
• Functional module discovery: Discovering unknown protein complexes and biological pathways from PPI data for downstream experimental validation.

Methodology:

Data integration of protein gene ontology attributes with network topology; adaptive aggregation of structural information from low- and high-order graphs based on node distribution and graph node smoothness; module prediction via core cliques and expansion algorithms.