Overall protein structure quality assessment using hydrogen-bonding parameters

Area of Science:

• Structural Biology
• Biophysics
• Computational Biology

Background:

• Atomic model refinement at low resolution presents challenges due to insufficient experimental data detail.
• Current refinement practices often employ Ramachandran plots and rotameric states, potentially diminishing their validation power.
• There is a need for novel model-validation criteria not typically used as refinement targets.

Purpose of the Study:

• To investigate the utility of hydrogen bond geometry as a novel criterion for atomic model validation.
• To establish a method for assessing atomic model quality using conserved hydrogen bond distributions.

Main Methods:

• Systematic analysis of hydrogen bond geometries in high-resolution protein models from the Protein Data Bank.
• Characterization of hydrogen donor and acceptor atom distributions and their geometric properties.
• Development of a validation approach based on observed hydrogen bond geometric patterns.

Main Results:

• Analysis revealed distinct and conserved geometric distributions for hydrogen bonds in quality-filtered protein structures.
• These conserved geometric features can be quantified and utilized for assessing atomic model quality.
• The proposed method provides an independent validation criterion for atomic models.

Conclusions:

• Hydrogen bond geometry offers a valuable and underutilized resource for atomic model validation, particularly at low resolution.
• This approach enhances the reliability of atomic models without interfering with existing validation metrics.
• The findings contribute to more accurate protein structure determination and analysis.