Selection of high-producing clones by a relative titer predictive model using image analysis

Area of Science:

• Biotechnology
• Biopharmaceutical Manufacturing
• Cell Line Development

Background:

• Monoclonal antibodies (Mabs) drive biopharmaceutical industry growth, necessitating efficient cell line development.
• Traditional clone selection is laborious, involving extensive evaluation of cell growth, density, titer, and product quality.
• Optimizing cell line development is crucial for rapid, consistent, and cost-effective biologics production.

Purpose of the Study:

• To develop an efficient clone selection strategy for biopharmaceutical companies.
• To create a predictive model for estimating relative titer (RT) using quantitative image analysis.
• To improve the speed and accuracy of identifying high-producing cell clones.

Main Methods:

• Developed a relative titer (RT) prediction model utilizing quantitative information from microscope images.
• Extracted data during the cell line development process for model training and validation.
• Evaluated the RT prediction model's performance on 50 clones across 5 distinct cell lines.

Main Results:

• The RT prediction model successfully identified high-producing clones when using the same host cells.
• Model predictions were less accurate when different host cells were involved.
• Demonstrated that quantitative image data from cell line development offers valuable insights.

Conclusions:

• Presented the first predictive model for estimating relative productivity of Chinese hamster ovaries (CHO) clones.
• The developed RT prediction model serves as a proof of concept for image-based cell line development.
• This study lays the groundwork for future predictive models to enhance clone selection in biologics manufacturing.