AIST Develops Novel Technology for Culture Medium Quality Assessment Using Machine Learning and Fluorescent Polymers: Revolutionizing Biomanufacturing

国立研究開発法人産業技術総合研究所 (AIST) Japan

Overview

The National Institute of Advanced Industrial Science and Technology (AIST) has developed a new “chemical tongue” sensor technology for assessing the quality of culture media and supplements in biomanufacturing. This system combines multiple fluorescent polymers with machine learning to accurately identify subtle compositional differences and degradation states through unique fluorescence patterns. By preventing culture-related issues, it contributes significantly to improving the quality of pharmaceuticals and regenerative medicine products.

In Depth

Importance of Culture Medium Quality Control in Biomanufacturing

In “biomanufacturing”—the production of pharmaceuticals, regenerative medicine products, and functional materials using cells or microorganisms—the cell culture process is one of the most critical factors determining product quality and production efficiency. The quality of culture media and supplements, which form the foundation of this cultivation, directly impacts cell growth and target substance production, thus requiring stringent quality control. However, conventional analytical methods have found it challenging to simply and rapidly assess subtle differences in media composition, batch-to-batch variability, or pre-culture degradation states.

AIST’s Developed “Chemical Tongue” Sensor Technology

To address this challenge, a research team at the National Institute of Advanced Industrial Science and Technology (AIST) has developed an innovative culture medium quality assessment technology. The core of this technology lies in a “chemical tongue” sensor that combines the following key elements:

• Multiple Fluorescent Polymers: Various types of fluorescent polymers are employed, each exhibiting unique fluorescence responses to different components or combinations of components within the culture medium. Particularly, the incorporation of polymers with “Aggregation-Induced Emission (AIE)” properties enables more sensitive detection of changes in molecular states within solutions.
• Fluorescence Patterns and Machine Learning: The complex fluorescence response patterns obtained from multiple fluorescent polymers are analyzed using machine learning algorithms. This approach allows for the integrated and highly accurate identification of the culture medium’s quality and state as a “chemical fingerprint,” rather than by detailed analysis of individual components.

This technology enables rapid and straightforward detection of subtle compositional differences and temporal changes in culture media that were previously often overlooked.

Industrial Applications and Future Outlook

This new technology promises to dramatically improve quality control in biomanufacturing. By accurately assessing culture medium quality before cultivation, it can prevent culture-related issues, significantly enhancing production stability and product quality uniformity. This represents a crucial advancement directly impacting the stable supply of pharmaceuticals and the safety assurance of regenerative medicine products. AIST plans to integrate this technology with more compact and portable detection devices for real-time quality control systems in manufacturing settings. Furthermore, validation for expanding its application range to various culture media and supplements is underway, with expectations that it will serve as a foundational technology strongly supporting the development of a sustainable biomanufacturing industry.