Biomanufacturing innovation relies on an efficient design of experiments (DoE) to optimize processes and product quality. Traditional DoE methods, ignoring the underlying bioprocessing mechanisms, often suffer from a lack of interpretability and sample efficiency. This limitation motivates us to create a new optimal learning approach that can guide a sequential DoEs for digital twin model calibration. In this study, we consider a multi-scale mechanistic model for cell culture process, also known as Biological Systems-of-Systems (Bio-SoS), as our digital twin. This model with modular design, composed of sub-models, allows us to integrate data across various production processes. To calibrate the Bio-SoS digital twin, we evaluate the mean squared error of model prediction and develop a computational approach to quantify the impact of parameter estimation error of individual sub-models on the prediction accuracy of digital twin, which can guide sample-efficient and interpretable DoEs.

翻译：生物制造创新依赖于高效的设计实验（DoE）以优化工艺与产品质量。传统DoE方法忽略底层生物加工机制，常面临可解释性不足和样本效率低下的问题。这一局限性促使我们开发一种新的最优学习方法，可引导用于数字孪生模型校准的序贯DoE。本研究中，我们采用细胞培养过程的多尺度机理模型——即生物系统之系统（Bio-SoS）——作为数字孪生载体。该模型采用模块化设计，由多个子模型构成，可整合不同生产过程中的数据。为校准Bio-SoS数字孪生，我们评估模型预测的均方误差，并开发计算方法量化各子模型参数估计误差对数字孪生预测精度的影响，从而指导样本高效且可解释的DoE。