Related Experiment Videos
Maximum-likelihood density modification using pattern recognition of structural motifs.
1Bioscience Division, Mail Stop M888, Los Alamos National Laboratory, Los Alamos, NM 87545, USA. terwilliger@lanl.gov
Acta Crystallographica. Section D, Biological Crystallography
|November 22, 2001
Summary
This study introduces a new likelihood-based pattern recognition method for electron density modification in crystallography. The approach significantly improves phase accuracy by recognizing structural motifs like helices, outperforming existing methods.
Area of Science:
- Crystallography
- Structural Biology
- Computational Chemistry
Background:
- Likelihood-based methods are crucial for electron density modification in macromolecular crystallography.
- Current methods like solvent flattening and histogram matching have limitations in phase improvement.
Purpose of the Study:
- To extend the likelihood-based approach to incorporate electron density pattern recognition.
- To improve phase accuracy in crystallographic maps by integrating prior knowledge of structural motifs.
Main Methods:
- Developed a likelihood-based framework that reformulates map likelihoods to include terms for recognized structural elements.
- Integrated terms for solvent flatness and protein electron density distribution.
- Tested the method by recognizing helical segments in a protein crystal structure.
Main Results:
- The pattern-recognition method demonstrated substantial phase improvement compared to conventional and likelihood-based solvent flattening and histogram matching.
- Successfully recognized helical segments, validating the approach.
Conclusions:
- The developed likelihood-based pattern-recognition method offers a significant advancement in electron density modification.
- This approach can be generalized to recognize and incorporate prior knowledge of various structural motifs for enhanced crystallographic phasing.