A picture is worth a thousand diffraction spots: using photometry and computer vision to perform rapid...

Area of Science:

• Crystallography
• Structural Biology
• Materials Science

Background:

• Serial crystallography is crucial for determining protein and material structures.
• Existing serial crystallography methods often require extensive sample preparation and have limitations in throughput.
• There is a need for more efficient and automatable methods for serial crystallography.

Purpose of the Study:

• To introduce and validate a novel 'photometric selection' approach for serial crystallography.
• To demonstrate the method's high throughput, sample tolerance, low cost, and automation potential.
• To showcase the broad applicability of photometric selection across different crystalline systems.

Main Methods:

• Utilizing image recognition algorithms from the computer vision project OpenCV for crystal identification.
• Loading and distributing crystalline samples onto a transparent substrate for automated analysis.
• Employing an in-line camera system for real-time crystal detection and selection.

Main Results:

• Photometric selection requires minimal sample refinement, unlike traditional serial techniques.
• High-quality datasets were successfully collected from diverse systems: organometallic, organic, and metal-organic frameworks.
• Data collection speed was enhanced, being up to six times faster than established grid-scanning techniques.

Conclusions:

• Photometric selection is a robust and versatile method for advancing serial crystallography.
• The technique significantly improves efficiency and reduces costs in crystallographic data collection.
• This approach has broad implications for structural determination across various scientific disciplines.