Depth determination in crystal lography using auto-focusing algorithms - biomed 2010

Area of Science:

• Crystallography
• Drug Discovery
• Robotics
• Machine Vision

Background:

• Automated crystallography is crucial for modern drug discovery but faces engineering challenges.
• The manual crystal harvesting step is a significant bottleneck in achieving full automation.
• Current methods for crystal harvesting rely on manual micro-manipulation and microscopy.

Purpose of the Study:

• To automate the crystal harvesting process in drug discovery.
• To develop a machine vision system for identifying crystals and determining their 3D coordinates.
• To investigate autofocusing algorithms for accurate crystal depth determination in microtiter plates.

Main Methods:

• Utilized machine vision to identify viable crystals and ascertain their 3D coordinates.
• Explored autofocusing algorithms, analogous to those in digital single-lens reflex (DSLR) cameras, for depth measurement.
• Compared autofocusing algorithms against complex methods like optical coherence tomography (OCT) and confocal reflectometry.
• Evaluated seven distinct autofocusing algorithms for their efficacy.

Main Results:

• Autofocusing algorithms show promise as a cost-effective alternative for determining crystal depth.
• The study successfully investigated the application of DSLR-like autofocusing techniques in crystallography.
• Results provide insights into the performance of different autofocusing algorithms for this specific application.

Conclusions:

• Autofocusing algorithms offer a viable and potentially less expensive method for automating crystal depth determination in crystallography.
• This automated approach can enhance the efficiency of robotic harvesting in drug discovery pipelines.
• Further development in machine vision and autofocusing can overcome key engineering obstacles in crystallography automation.