Related Experiment Videos
Inferring domain-domain interactions from protein-protein interactions
Minghua Deng1, Shipra Mehta, Fengzhu Sun
1Program in Molecular and Computational Biology, Department of Biological Sciences, University of Southern California, Los Angeles, California 90089, USA.
Genome Research
|October 9, 2002
Summary
Understanding protein interactions at the domain level provides a global view of cellular networks. This study infers domain-domain interactions from protein-protein interaction data, revealing conserved patterns and predicting novel interactions.
Area of Science:
- Molecular Biology
- Systems Biology
- Bioinformatics
Background:
- Protein-protein interactions are crucial for cellular functions.
- Understanding these interactions at the domain level offers insights into the broader protein interaction network.
- The yeast Saccharomyces cerevisiae proteome has been extensively studied for protein interactions.
Purpose of the Study:
- To infer interacting protein domains from large-scale protein-protein interaction data.
- To study conserved patterns of domain-domain interactions.
- To predict novel protein-protein interactions based on inferred domain interactions.
Main Methods:
- Utilized yeast two-hybrid assay data to identify protein-protein interactions.
- Employed Maximum Likelihood Estimation to infer domain-domain interactions using PFAM database domains.
- Validated predictions by comparing with existing protein-protein interaction databases and gene expression profiles.
Main Results:
- Successfully inferred probabilities for domain-domain interactions.
- Predicted protein-protein interactions showed significant overlap with known interactions from other methods.
- Predicted interaction pairs exhibited higher gene expression profile correlation than random pairs.
- Demonstrated robustness in handling incomplete data and experimental errors.
Conclusions:
- Inferring domain-domain interactions is a powerful approach to understand protein interaction networks.
- The method accurately predicts protein-protein interactions and identifies novel interactions.
- This approach provides a robust framework for analyzing complex biological interaction data.