Principal components analysis of protein structure ensembles calculated using NMR data

Area of Science:

• Structural biology
• Computational biophysics
• Biomolecular NMR spectroscopy

Background:

• Distinguishing converged structures from outliers is crucial for biomolecular structure calculation using NMR data.
• Current methods may struggle to automatically classify structural ensembles and identify dynamic regions.

Purpose of the Study:

• To introduce and validate Principal Components Analysis (PCA) as a method for automated classification of NMR-derived biomolecular structures.
• To demonstrate PCA's capability in identifying correlated structural variations and highlighting flexible regions within protein structures.

Main Methods:

• Protein structures were represented in reduced complexity forms suitable for PCA.
• Two distinct structural representations were developed and compared.
• PCA was applied to an ensemble of calculated structures for a 28 amino acid peptide.

Main Results:

• Both structural representations yielded equivalent and accurate PCA results.
• PCA successfully classified structures even when the ensemble contained two distinct protein conformations.
• The analysis effectively identified specific regions of structural variation and differences between conformations.

Conclusions:

• PCA offers a robust and automated approach for classifying structural ensembles derived from NMR data.
• This method aids in identifying reliable structures and understanding protein dynamics.
• PCA is effective even in the presence of multiple conformations within the structural ensemble.