TSEMA

TSEMA predicts protein-protein interactions by comparing phylogenetic trees of interacting protein families to detect co-evolutionary signals that indicate functional relationships.

Key Features:

• Phylogenetic Tree Comparison: Compares phylogenetic trees derived from sequences of associated protein families and identifies mappings between family members by maximizing tree similarity.
• Monte Carlo Approach: Employs a Monte Carlo heuristic to explore the combinatorial mapping space and generate initial mappings for large protein families.
• Large-Scale Prediction Capability: Scales to extensive datasets, for example applied to over 67,000 E. coli protein pairs to predict 2,742 interacting pairs.
• Statistical Evaluation and Validation: Incorporates statistical evaluation mechanisms to assess predictive performance across test sets.

Scientific Applications:

• Proteomics: Infers protein-protein interactions to support reconstruction of cellular interaction networks.
• Computational Biology: Analyzes co-evolutionary relationships between protein families to study evolutionary interdependencies.
• Functional Genomics: Supports functional annotation by linking proteins through predicted interactions.
• Drug Discovery: Identifies candidate interacting proteins that may inform target validation and interaction-based drug studies.

Methodology:

Constructs phylogenetic trees from protein family sequences, uses Monte Carlo simulations to propose mappings that maximize tree similarity, and applies statistical evaluation to assess predictive capacity.