Related Experiment Videos
Exploiting the co-evolution of interacting proteins to discover interaction specificity
Arun K Ramani1, Edward M Marcotte
1Institute for Cellular and Molecular Biology, Center for Computational Biology and Bioinformatics, University of Texas at Austin, Austin, TX 78712, USA.
Journal of Molecular Biology
|March 5, 2003
Summary
Predicting protein interaction specificity is now possible using evolutionary relationships. New computational methods align phylogenetic trees of protein families to identify specific interaction partners, improving cellular function understanding.
Area of Science:
- Biochemistry
- Computational Biology
- Evolutionary Biology
Background:
- Protein interactions are crucial for cellular functions.
- Predicting specific interactions between protein families (e.g., ligands and receptors) remains a challenge.
- High-throughput methods are needed to map these interactions.
Purpose of the Study:
- To develop and validate computational methods for predicting protein interaction specificity.
- To leverage evolutionary relationships within protein families for prediction.
- To understand the evolutionary processes driving interaction specificity.
Main Methods:
- Matrix alignment: Aligning protein family similarity matrices to find optimal relationships.
- 3D embedding: Visualizing interacting protein families in spatial representations.
- Phylogenetic tree alignment: Aligning evolutionary trees of interacting families to define partners.
Main Results:
- Successfully predicted protein interaction specificities for over 18 protein families.
- Prediction accuracy correlates with phylogenetic tree complexity.
- Developed a model for the co-evolution of interacting protein families.
Conclusions:
- Evolutionary relationships within protein families can accurately predict physical interaction specificities.
- Computational alignment of phylogenetic trees offers a powerful approach to defining specific protein partners.
- The findings provide insights into the evolution of protein interaction networks.