Back

Advancing Biomanufacturing Excellence Through Artificial Intelligence

Discover the role of Artificial Intelligence in revolutionizing biologics manufacturing and learn how Cytovance and partners at the University of Oklahoma are contributing to setting the new standard.

Source: Cytovance Biologics

 Artificial Intelligence (AI) in Biomanufacturing
At its core, AI involves the development of sophisticated algorithms and computational systems that can analyze vast sets of data, identify patterns and insights, and then leverage that knowledge to make predictions, decisions, and take actions, all with speed and accuracy that can exceed human capabilities. The broad application of AI across a diverse list of industries is revolutionizing how we approach complex challenges. In healthcare, AI-powered tools are accelerating drug discovery by rapidly screening millions of chemical compounds to identify promising drug candidates, assisting clinicians in making more accurate diagnoses and treatment plans for patients, and now taking the first steps in transforming biomanufacturing.

The application of AI to biomanufacturing is particularly exciting, where the technology is being leveraged to optimize the complex biological processes involved in producing life-saving therapeutics and vaccines. By harnessing the pattern recognition and decision-making capabilities of AI, scientists and engineers can fine-tune these delicate procedures and controls, improving yields, detecting defects, and ensuring the highest levels of quality and consistency. As AI continues to advance, we can expect to see its impact expand, empowering us to tackle even the most challenging hurdles with greater efficiency, accuracy, and insight than ever before.

At Cytovance, we are at the forefront of fostering industry innovation and are collaborating with key researchers to use AI to develop a model that would further optimize biomanufacturing capabilities beyond what is currently possible.

 

Current Contributions to AI Development

Recently, Cytovance collaborated with University of Oklahoma (OU) researchers to produce data for “An imbalance-aware BiLSTM for control chart patterns early detection”  published in Expert Systems with Applications journal and “Optimizing Protein Titer Production using Animal Cells: Predictive Modeling and Recommendations for Enhanced Yield,”  presented at the 2024 Institute of Industrial and Systems Engineers (IISE) Annual Meeting, that leveraged our extensive in-house datasets from Chinese Hamster Ovary (CHO)-based process development work across multiple cell lines. By tapping into this wealth of data, the collaborators were able to identify a number of significant parameters that can have a major impact on protein titer production. This comprehensive analysis examined a variety of environmental factors, feed strategies, and cell line characteristics to uncover the key drivers of productivity. Building on these insights, advanced machine learning-based analytical models were then developed that were capable of accurately predicting protein titer outputs. Through in-depth analysis of these predictive models, complex relationships were able to be dissected between the various input variables and the final titer production.

Looking forward, the practical implications of this work are substantial, as these modeling capabilities can be applied across the broader biopharmaceutical and biotechnology industries to drive increased productivity and cost-effectiveness in mammalian-based protein production processes. With our academic collaborators, we envision further refining these predictive models by incorporating additional real-world datasets, exploring alternative machine-learning algorithms, and conducting deeper investigations into the interplay between the identified significant variables. This work also emphasizes the value of continued collaboration with academic and industry partners to validate these models in practical settings and explore their scalability to ensure maximum impact and long-term benefits to patients.

The success of this CHO-focused approach has inspired Cytovance and partners at OU to explore applying similar machine learning-driven strategies to microbial-based bioprocesses, further expanding the reach and utility of these cutting-edge analytical capabilities. Leveraging Cytovance’s experience and extensive datasets as a leading microbial CDMO, and in collaboration with customers eager to contribute to the future of biomanufacturing, we have begun examining the key parameters that can be optimized for E. coli fermentation, such as pH, temperature, and oxygen levels to determine the conditions to achieve the highest product titers in a process. This will provide valuable insights into the standard processes that are used in microbial-based biomanufacturing. The findings from this research have been submitted to a peer-reviewed conference. Both teams look forward to the data offering the biopharmaceutical community a deeper understanding of how machine learning can be harnessed to enhance and streamline E. coli-based programs.

 

The Future of AI in Biopharmaceutical Production

Outside of perfecting processes, the use of AI promises to revolutionize the field of strain and cell line development for biopharmaceutical production. By leveraging advanced AI algorithms, researchers could better genetically engineer highly optimized strains and cell lines that are tailored to maximize titer potential. Beyond strain development, AI is also poised to transform bioreactor design through the use of predictive modeling. AI-driven bioreactor designs can simulate complex biological and engineering factors to predict optimal operating conditions, leading to more efficient and productive fermentation processes. This predictive modeling can extend to the automation of control systems as well, enabling consistent, high-quality production even at the largest manufacturing scales. By integrating these AI-powered advancements with other emerging bioprocessing technologies, biopharmaceutical companies can establish highly streamlined, data-driven production workflows.

By supplying data and taking an active role in vetting an AI approach to optimize critical biomanufacturing variables, Cytovance is on track to unlock new levels of efficiency and productivity for our customers. Leveraging advanced analytics and machine learning, we can streamline development plans and zero in on the two to three most critical parameters to focus our optimization efforts. This targeted approach can save substantial time and resources compared to a more traditional empirical methodology. Additionally, the insights gleaned from the AI analysis can simplify the later process characterization stage, helping to easily identify the key process controls necessary to maintain a robust, optimized production process throughout the entirety of manufacturing.

Ultimately, the power of AI to accelerate and refine biomanufacturing holds immense promise, paving the way for continued advancements and innovations in this dynamic field. Companies at the forefront of these efforts, like Cytovance, are set to drive the biomanufacturing industry forward through the combination of real-world data and collaborations with academic institutions, ushering in a new era of increased efficiency, reduced costs, and heightened product quality that serve patients better in the long run.

—

Talayeh Razzaghi, Ph.D. University of Oklahoma

Contributing Author: Talayeh Razzaghi, Ph.D.

Assistant Professor of School of Industrial and Systems Engineering

Assistant Professor of Data Science and Analytics Institute

Gallogly College of Engineering, University of Oklahoma

We would like to extend a special thank you to  Dr. Talayeh Razzaghi and her team, including Rachel Bennett, Grant Benson, and Mohammad Derakhshi, for inviting us to contribute to their research.

—