Automatic domain decomposition of proteins by a Gaussian Network Model

Area of Science:

• Structural biology
• Computational biophysics
• Bioinformatics

Background:

• Proteins are composed of independent folding units called domains.
• Defining and classifying these domains is crucial for understanding protein function.
• Existing methods for domain classification can be subjective or computationally intensive.

Purpose of the Study:

• To develop an automated method for classifying protein structural domains.
• To adapt the Gaussian Network Model (GNM) for domain classification.
• To validate the GNM-based domain classification against expert assignments.

Main Methods:

• Utilized the Gaussian Network Model (GNM) to analyze protein dynamics.
• Adapted GNM to identify substructures with independent motion.
• Applied the algorithm to a dataset of 55 non-redundant proteins.
• Compared computational domain assignments with crystallographer visual assignments.

Main Results:

• The GNM successfully identified and classified protein domains based on motion.
• The automated classification showed strong agreement with expert visual assignments.
• The method demonstrated its ability to decompose proteins into structurally and dynamically distinct units.

Conclusions:

• The Gaussian Network Model (GNM) provides an effective computational approach for automated protein domain classification.
• This method offers an objective and efficient alternative to manual domain assignment.
• The algorithm has broad applicability for decomposing large macromolecular systems into motionally decoupled subsystems.