Machine learning‐based model predictive controller design for cell culture processes

Abstract

The biopharmaceutical industry continuously seeks to optimize the critical quality attributes to maintain the reliability and cost‐effectiveness of its products. Such optimization demands a scalable and optimal control strategy to meet the process constraints and objectives. This work uses a model predictive controller (MPC) to compute an optimal feeding strategy leading to maximized cell growth and metabolite production in fed‐batch cell culture processes. The lack of high‐fidelity physics‐based models and the high complexity of cell culture processes motivated us to use machine learning algorithms in the forecast model to aid our development. We took advantage of linear regression, the Gaussian process and neural network models in the MPC design to maximize the daily protein production for each batch. The control scheme of the cell culture process solves an optimization problem while maintaining all metabolites and cell culture process variables within the specification. The linear and nonlinear models are developed based on real cell culture process data, and the performance of the designed controllers is evaluated by running several real‐time experiments.

Cell culture product quality attribute prediction using convolutional neural networks and Raman spectroscopy

“…With the advent of process analytical technology (PAT), various tools have been implemented to monitor quality attributes of biopharmaceutical processes (FDA, 2004; Wasalathanthri et al, 2020). Recently, model‐based advanced process control techniques have been adopted to both improve yield productivity and reduce waste in cell culture processes (Rashedi et al, 2022, 2023). High variability and nonlinear behavior together with the lack of frequent process measurements and the small amount of available data are the challenges to process optimization which demands adaptive modeling and control of cell culture processes (Craven et al, 2014).…”

“…The current method for controlling cell culture variables in the industry is accomplished either manually by lab operators or using traditional proportional‐integral‐derivative (PID) controllers through addition of bolus feeds (Mehdizadeh et al, 2015). Recent works have proved that the protein production in the cell culture process can be improved through the implementation of advanced control strategies such as model predictive control (Craven et al, 2014; Rashedi et al, 2022, 2023); however, these controllers may suffer from low modeling accuracy due to infrequent sampling. To tackle this challenge and improve the controller performances, automated and frequent estimation of quality attributes using noninvasive in‐situ methods such as Raman (Abu‐Absi et al, 2011) or near‐infrared (Cervera et al, 2009) spectroscopy are necessary.…”