Qualitative and Quantitative Protein Complex Prediction Through Proteome-Wide Simulations.

Area of Science:

• Proteomics and Systems Biology
• Computational Biology and Bioinformatics

Background:

• Identifying protein complexes is crucial for understanding cellular mechanisms and drug development.
• Current algorithms for protein complex prediction are limited to qualitative assessments, lacking quantitative abundance data.

Purpose of the Study:

• To develop a novel computational approach for predicting both the composition and abundance of protein complexes.
• To address the limitations of existing methods by providing quantitative predictions.

Main Methods:

• Developed Simulation based Complex Prediction (SiComPre), a method utilizing stochastic simulations.
• Integrated multi-source data including protein abundances, domain-domain interactions, and functional annotations.
• Applied the method to predict yeast and human protein complexes and their abundances.

Main Results:

• SiComPre demonstrates superior qualitative prediction accuracy for yeast and human protein complexes compared to existing methods.
• This is the first method capable of predicting the quantitative abundance of protein complexes.
• Successfully applied SiComPre to predict complexome alterations following bortezomib drug treatment.

Conclusions:

• SiComPre offers the first quantitative predictions of molecular complex abundance, alongside state-of-the-art qualitative predictions.
• The method is adaptable for predicting tissue-specific and condition-dependent complexome variations.
• Enables deeper insights into cellular regulation and potential therapeutic targets.