A novel method for identifying disease associated protein complexes based on functional similarity protein complex...

Area of Science:

• Biochemistry and Molecular Biology
• Computational Biology
• Genomics and Proteomics

Background:

• Protein complexes, crucial for cellular functions, are increasingly identified via computational and purification methods.
• However, their direct roles in disease etiology remain underexplored, with limited studies focusing on disease-associated protein complexes.
• Existing methods often rely on outdated data and complex networks, limiting their applicability.

Purpose of the Study:

• To propose a novel computational method for identifying disease-protein complex associations.
• To develop a framework for constructing functional similarity protein complex networks.
• To introduce an efficient algorithm for ranking candidate disease-protein complexes.

Main Methods:

• Constructed functional similarity protein complex networks based on shared proteins, GO terms, or protein interactions.
• Employed a neighborhood-based algorithm for local similarity measurement to rank potential disease-protein complex associations.
• Validated the method against state-of-the-art network propagation algorithms.

Main Results:

• The proposed method significantly outperformed existing algorithms in predictive performance across various network constructions.
• The algorithm demonstrated a substantial speed improvement, running approximately 32 times faster than comparative methods.
• High AUC values were consistently achieved, irrespective of network construction or algorithm choice, outperforming complex heterogeneous network approaches.
• Application to prostate cancer identified 69 validated candidate protein complexes among the top 100 ranked associations.

Conclusions:

• The developed framework for functional similarity protein complex networks and the associated neighborhood-based algorithm effectively identify novel disease-protein complex associations.
• This method offers a more efficient and accurate approach for uncovering the links between protein complexes and human diseases.
• The findings suggest a promising avenue for future research in disease mechanism and biomarker discovery.