Modeling protein assemblies in the proteome.

Area of Science:

• Structural biology
• Computational biology
• Biophysics

Background:

• Proteins function in multimolecular assemblies, making their atomic-scale structure determination crucial.
• Current computational methods for protein assembly structure prediction have limitations, including high computational cost, inability to handle large-scale studies, and restrictions to symmetrical complexes.

Purpose of the Study:

• To develop a novel computational strategy for constructing protein assemblies.
• To overcome limitations of existing methods by enabling the prediction of homo-/hetero-complexes, including symmetric and asymmetric arrangements, without component number limitations.

Main Methods:

• Utilized a motif-based protein interaction prediction tool, PRISM (Protein Interactions by Structural Matching), to predict binary interactions.
• Developed a computational strategy to build multimolecular assemblies based on predicted pairwise interactions.
• Incorporated conformational changes and leveraged electron microscopy density maps for solution selection.

Main Results:

• Successfully constructed homo-/hetero-complexes and symmetric/asymmetric complexes computationally.
• The method is applicable to large-scale studies and considers conformational changes.
• Demonstrated the ability to select accurate solutions using experimental data (electron microscopy density maps).

Conclusions:

• The developed method offers a significant advancement in computationally predicting protein assembly structures.
• This approach addresses limitations of previous methods, enabling broader application in structural biology and proteome studies.
• The strategy provides a foundation for understanding complex cellular machinery at an atomic level.