Efficient merging of data from multiple samples for determination of anomalous substructure

Area of Science:

• Crystallography
• Structural Biology
• Biophysics

Background:

• Merging data from multiple crystals is crucial for determining atomic substructures, particularly with weak anomalous scatterers or poor diffraction.
• Assessing sample isomorphism using unit-cell parameters, anomalous signal correlation, and data similarity is key for successful data merging.

Purpose of the Study:

• To evaluate an efficient protocol for merging data from numerous samples using local scaling and anomalous signal optimization.
• To assess the effectiveness of this protocol in handling non-isomorphism and radiation damage in crystallographic data.

Main Methods:

• Implementation of local scaling, anomalous signal optimization, and data-set weighting in the phenix.scale_and_merge software.
• Phasing by single-wavelength anomalous diffraction (SAD) of native sulfur atoms in the protein NS1, using data from 28 samples.
• Assessment of merged data quality by reducing multiplicity through exclusion of individual crystals or radiation-damaged data segments.

Main Results:

• The local-scaling and anomalous-optimization protocol yielded merged datasets with superior anomalous signal quality indicators compared to standard global-scaling methods.
• Local-scaled data demonstrated higher success rates in substructure determination.
• The protocol effectively managed sample non-isomorphism and radiation-induced decay, with equivalent anomalous signal at comparable multiplicities.

Conclusions:

• The developed protocol offers an efficient method for merging crystallographic data from multiple samples, improving anomalous signal quality and substructure determination.
• This approach robustly handles common challenges in data collection, such as sample non-isomorphism and radiation damage.
• Data quality and structure determination success are strongly correlated with anomalous signal metrics and data multiplicity.