EUVIMEDOdkrijte EUVIMED
Open access · CC BY 4.0

LLM‐Guided Parameter Optimization for Mechanistic CHO Cell Bioreactor Models

Han Bit Kim iD, Janghan Lee iD, Seo-Yeon Kim iD, Jihoon Kim iD, Moo Sun Hong iD

Biotechnology and Bioengineering · 2026-07-01

DOI: 10.1002/bit.70300 ↗

At a glance

Journal article in Biotechnology and Bioengineering, published online on 2026-07-01. Written by Han Bit Kim and 4 co-authors. Institution: Department of Chemical and Biological Engineering, Seoul National University, Seoul, Republic of Korea. Major MEDLINE topics: Bioreactors, Models, Biological. 16 references. The full text is freely available (open access, CC BY 4.0). Funded by: Seoul National University, Ministry of Trade, Industry and Resources (MOTIR, Korea).

PovzetekSubjectsBibliografski podatkiCiteRelated studies

Abstract (original language)

Mechanistic models of Chinese Hamster Ovary (CHO) cell cultures are essential for efficient bioprocess development and control, yet parameter estimation remains a critical challenge in model calibration. As model complexity increases, the number of parameters grows and their values become highly system‐specific due to biological variability across cell lines and operating conditions. Consequently, estimation requires broad search spaces, leading to inefficient exploration and increased risk of entrapment in local minima during numerical optimization. To address this, a structure‐aware optimization framework is proposed that integrates large language model (LLM) reasoning into the parameter estimation loop. The LLM analyzes governing equations and trajectory discrepancies to recommend directional adjustments for parameter boundaries. These insights are combined with correlation‐based scaling to adaptively refine the search space, which is explored using stochastic global search methods (Latin hypercube sampling [LHS] and Covariance matrix adaptation evolutionary strategy [CMA‐ES]). Applied to fed‐batch CHO cell culture data, the framework escapes local minima and resolves complex, nonlinear metabolite dynamics, including lactate and ammonium shifts. Statistical validation across multiple batches and random seeds shows consistent reduction in prediction error and improved optimization robustness. This study demonstrates the potential of integrating structure‐aware LLM guidance with numerical optimization to improve the reliability and efficiency of bioprocess modeling.

Subjects

MeSH terms (MEDLINE)

BioreactorsModels, BiologicalAnimalsCHO CellsCricetulusLarge Language ModelsCricetinae

Author keywords

CHO cellsbioreactor modelinglarge language models (LLMs)mechanistic modelingparameter estimation

Authors and institutions

  • Department of Chemical and Biological Engineering, Seoul National University, Seoul, Republic of Korea.
  • New Drug Development Center, Osong Medical Innovation Foundation, Cheongju, Republic of Korea.
  • College of Pharmacy, Chungbuk National University, Cheongju, Republic of Korea.

ORCID profiles: Han Bit Kim, Janghan Lee, Seo-Yeon Kim, Jihoon Kim, Moo Sun Hong

Funding

  • Seoul National University
  • Ministry of Trade, Industry and Resources (MOTIR, Korea)

Cite

APA

Kim, H. B., Lee, J., Kim, S. Y., Kim, J., & Hong, M. S. (2026). LLM‐Guided Parameter Optimization for Mechanistic CHO Cell Bioreactor Models. Biotechnology and Bioengineering. https://doi.org/10.1002/bit.70300

Vancouver

Kim HB, Lee J, Kim SY, Kim J, Hong MS. LLM‐Guided Parameter Optimization for Mechanistic CHO Cell Bioreactor Models. Biotechnology and Bioengineering. 2026. doi:10.1002/bit.70300

BibTeX
@article{Kim2026,
  author = {Kim, Han Bit and Lee, Janghan and Kim, Seo-Yeon and Kim, Jihoon and Hong, Moo Sun},
  title = {LLM‐Guided Parameter Optimization for Mechanistic CHO Cell Bioreactor Models},
  journal = {Biotechnology and Bioengineering},
  year = {2026},
  doi = {10.1002/bit.70300}
}
Start for free →