PFP-GO

PFP-GO predicts protein functions by integrating protein sequences, domain annotations, and protein-protein interaction networks to rank and propagate Gene Ontology (GO) terms to target proteins.

Key Features:

• Integration of Heterogeneous Data Sources: Processes protein sequences, domain annotations, and protein-protein interactions separately to leverage sequence, structural/functional domain, and network topology information.
• Ranking and Propagation of GO Terms: Ranks individual Gene Ontology (GO) terms based on relevance derived from the integrated datasets and propagates the ranked GO terms to target proteins.
• Performance Metrics: Evaluated using Precision (0.67), Recall (0.58), and F-Score (0.62).
• Validation through Multilayer Network Propagation: Validates top-ranked GO terms using multilayer network propagation techniques, including GO terms influencing the three-dimensional structure of the genome.

Scientific Applications:

• Genomics: Supports protein function annotation relevant to genome organization and interpretation.
• Proteomics: Facilitates functional annotation and interpretation of proteomic datasets.
• Systems Biology: Enables systems-level analysis by integrating interaction networks and functional annotations to elucidate molecular processes.

Methodology:

Each source—protein sequences, domain annotations, and interaction networks—is processed independently to extract relevant information.
Extracted information is used to rank GO terms for each protein based on functional relevance.
Ranked GO terms are propagated to target proteins.