<scp>MOO</scp> ‐based implementation of process analytical technology framework: Protein refolding as a case study

Abstract Protein refolding is usually lengthy, inefficient, costly, cumbersome, and prone to variations in end product quality. In this study, we propose two process analytical technology (PAT) based schemes for real‐time monitoring and control of a protein refolding process. The first scheme comprises a multiple partial least squares (MPLS) statistical process control (SPC) chart and a one‐dimensional convolutional neural network (1D‐CNN)‐powered Fourier transform infrared (FTIR) based analysis. The second scheme comprises the Shewhart SPC chart, a 1D‐CNN‐powered FTIR tool, and the non‐dominated sorting genetic algorithm II (NSGA‐II). When evaluated by inducing parameter deviations during refolding, the multi‐objective optimization (MOO)‐based PAT scheme demonstrated the ability to successfully identify these deviations and autonomously adjust process parameters, thereby ensuring that the final yield consistently remained within the specified target limits. The results of this study demonstrate that the NSGA‐II‐powered PAT system significantly enhances the consistency and reliability of the refolding reaction step.