UNIPred-Web

UNIPred-Web performs automated protein function prediction by integrating multiple biomolecular networks with a function-specific extension of the UNIPred algorithm that accounts for annotation imbalance.

Key Features:

• Integration of Biomolecular Networks: Integrates co-expression, shared domains, and physical interaction networks in a function-specific manner that addresses protein annotation imbalance.
• Consensus Network Generation: Produces consensus networks that aggregate evidence across input network types for downstream analysis.
• Visualization Capabilities: Provides visual representations of network topology to support exploration of protein relationships and consensus results.
• Vertex-centric Analysis: Implements a vertex-centric approach for detailed analysis of individual proteins and sub-networks.
• Support for Multiple Organisms: Supports biomolecular networks from nine different organisms.
• Custom Network Support: Accepts user-provided protein networks for integration and analysis.
• Gene Ontology Prediction: Predicts thousands of Gene Ontology (GO) protein functions.

Scientific Applications:

• Automated Function Annotation: Enables large-scale prediction and annotation of protein functions across thousands of GO terms.
• Hypothesis Generation for Unannotated Proteins: Facilitates generation of hypotheses about functions of previously unannotated proteins using integrated network evidence.
• Cross‑network and Cross‑species Analysis: Supports comparative analyses that combine multiple network types and data from nine organisms.

Methodology:

Extends the UNIPred algorithm to perform function-specific integration of biomolecular networks (co-expression, shared domains, physical interactions), accounts for annotation imbalance, generates consensus networks, and applies a vertex-centric approach for protein and sub-network analysis.