Protocol using similarity score and improved shrink-wrap algorithm for better convergence of phase-retrieval...

Area of Science:

• Crystallography and Materials Science
• Computational Imaging and Data Analysis

Background:

• X-ray diffraction imaging (XDI) reconstructs particle electron density maps using phase-retrieval (PR) algorithms.
• PR calculations can fail to produce realistic maps due to missing low-angle data and noise, hindering XDI analyses.
• Current PR methods present a computational bottleneck for accurate structure determination.

Purpose of the Study:

• To develop an efficient protocol for generating realistic electron density maps in XDI.
• To overcome limitations in PR calculations caused by missing low-angle scattering information and noise.
• To enhance the reliability and speed of XDI for non-crystalline particle structure analysis.

Main Methods:

• Proposed a protocol leveraging the empirical observation that realistic maps correlate with low similarity scores.
• Implemented concurrent PR calculations, modifying maps based on those with low similarity scores.
• Introduced a novel protocol for estimating particle shape to complement PR.
• Tested the protocol on diffraction patterns from colloidal gold particle aggregates.

Main Results:

• The proposed protocol significantly improved the probability of obtaining realistic electron density maps.
• The method demonstrated superior performance compared to standard PR calculations without the protocol.
• The protocol effectively addressed issues related to missing low-angle data and noise.
• Successful application to diffraction patterns from various colloidal gold particle aggregates.

Conclusions:

• The developed protocol efficiently yields realistic electron density maps for XDI.
• This approach reduces computational costs and accelerates XDI structure analysis.
• The protocol enables efficient XDI of non-crystalline particles using synchrotron X-rays and XFELs.