Abstract:Lay summary: This study examined the potential of machine learning to assist in bioreactor scale‐up. The findings demonstrated the capability of these algorithms to uncover complex non‐linear relationships among scale‐sensitive features, transfer knowledge, and predict process performance across scales. A method for predicting scaling factors for equivalent performance across scales was also developed and the characteristics of ideal datasets for future application of machine learning to scaling described. Bioreactor scale‐up and scale‐down have always been a topical issue for the biopharmaceutical industry and despite considerable effort, the identification of a fail‐safe strategy for bioprocess development across scales remains a challenge. With the ubiquitous growth of digital transformation technologies, new scaling methods based on computer models may enable more effective scaling. This study aimed to evaluate the potential application of machine learning (ML) algorithms for bioreactor scale‐up, with a specific focus on the prediction of scaling parameters. Factors critical to the development of such models were identified and data for bioreactor scale‐up studies involving CHO cell‐generated mAb products collated from the literature and public sources for the development of unsupervised and supervised ML models. Comparison of bioreactor performance across scales identified similarities between the different processes and primary differences between small‐ and large‐scale bioreactors. A series of three case studies were developed to assess the relationship between cell growth and scale‐sensitive bioreactor features. An embedding layer improved the capability of artificial neural network models to predict cell growth at a large‐scale, as this approach captured similarities between the processes. Further models constructed to predict scaling parameters demonstrated how ML models may be applied to assist the scaling process. The development of data sets that include more characterization data with greater variability under different gassing and agitation regimes will also assist the future development of ML tools for bioreactor scaling.

A Deep Learning Approach to Optimize Recombinant Protein Production in Escherichia coli Fermentations

PERISCOPE-Opt: Machine learning-based prediction of optimal fermentation conditions and yields of recombinant periplasmic protein expressed in <i>Escherichia coli</i>

Optimization of Phenyllactic Acid Biosynthesis and Separation by Machine Learning with Neural Network and Overlay Sampling Uniform Design

A machine learning‐based approach for improving plasmid DNA production in Escherichia coli fed‐batch fermentations

A perspective‐driven and technical evaluation of machine learning in bioreactor scale‐up: A case‐study for potential model developments

Genetic algorithm‐based semisupervised convolutional neural network for real‐time monitoring of Escherichia coli fermentation of recombinant protein production using a Raman sensor

Application Of Mechanistic Modelling And Machine Learning For Cream Cheese Fermentationphprediction

Towards improved penicillin fermentation via artificial neural networks

Fermentation Process Quality Prediction Using Teacher Student Stacked Sparse Recurrent Autoencoder

Modelling and Optimization of Fed-Batch Fermentation Processes Using Dynamic Neural Networks and Genetic Algorithms

Machine Learning Optimization of Photosynthetic Microbe Cultivation and Recombinant Protein Production

Data-Driven Soft Sensor Model Based on Deep Learning for Quality Prediction of Industrial Processes

Rolling Learning-Prediction Of Product Formation In Bioprocesses

MPEPE, a predictive approach to improve protein expression in E. coli based on deep learning

Hybrid deep modeling of a CHO-K1 fed-batch process: combining first-principles with deep neural networks

Optimizing Bioprocessing Efficiency with OptFed: Dynamic Nonlinear Modeling Improves Product-to-Biomass Yield

A deep learning-aided multi-objective optimization of a downstream process for production of monoclonal antibody products

Machine Learning algorithms for in-line monitoring during Yeast Fermentations based on Raman Spectroscopy

Deep hybrid modeling of a HEK293 process: Combining long short‐term memory networks with first principles equations

Low-N protein engineering with data-efficient deep learning

Deep model predictive control of gene expression in thousands of single cells