Determining protein structure from electron-density maps using pattern matching

Area of Science:

• Structural biology
• Computational biology
• Biophysics

Background:

• Protein structure determination is crucial for understanding biological function.
• Interpreting electron-density maps to build atomic models is a time-consuming bottleneck.
• Automated methods are needed to accelerate the process of protein structure elucidation.

Purpose of the Study:

• To develop an automated system, TEXTAL, for building protein structures from electron-density maps.
• To evaluate the accuracy and efficiency of TEXTAL in protein model generation.
• To present a novel computational approach for accelerating protein structure determination.

Main Methods:

• TEXTAL employs pattern recognition to identify similar regions between unknown electron-density maps and a database of known structures.
• Rotation-invariant features are extracted from spherical regions in the electron-density maps.
• Similarity is assessed using feature value differences and electron-density correlation coefficients.

Main Results:

• TEXTAL successfully built protein structures from various test electron-density maps.
• The system can automatically model entire protein structures within hours on a standard workstation.
• Models achieved high accuracy, with root-mean-square deviations of 0.6-0.7 Å (assuming C(alpha) positions).

Conclusions:

• TEXTAL offers a new, automated approach to protein structure determination.
• The system has the potential to significantly reduce the time and effort required for interpreting electron-density maps.
• TEXTAL facilitates the rapid generation of accurate protein models, aiding structural biology research.