High-Throughput Screening to Obtain Crystal Hits for Protein Crystallography

Area of Science:

• Structural Biology
• Biochemistry
• Crystallography

Background:

• X-ray crystallography is vital for determining macromolecular structures.
• Protein crystallization is a challenging, labor-intensive bottleneck in structural biology.
• Resource-limited institutions face significant hurdles in protein crystallization experiments.

Purpose of the Study:

• To present reproducible methods for facilitating protein crystal growth.
• To describe an automated high-throughput crystallization system.
• To detail an AI-driven approach for analyzing crystallization outcomes.

Main Methods:

• Utilized a 1,536-well microbatch-under-oil plate setup for broad parameter sampling.
• Employed state-of-the-art imaging modalities for monitoring crystal growth over 6 weeks.
• Implemented an artificial intelligence scoring algorithm for hit identification and an open-source interface for image analysis.

Main Results:

• Developed a reproducible workflow for protein crystallization screening.
• Demonstrated the efficacy of automated imaging and AI for distinguishing crystal hits.
• Streamlined the analysis of crystallization images, reducing manual effort.

Conclusions:

• The described high-throughput methods enhance reproducibility in protein crystallization.
• Automated systems and AI significantly improve the efficiency and accuracy of hit identification.
• This approach increases the likelihood of successful crystallization for structural studies.