Exploring the potential of machine learning for more efficient development and production of biopharmaceuticals

Abstract

Principles of Industry 4.0 direct us to predict how pharmaceutical operations and regulations may exist with automation, digitization, artificial intelligence (AI), and real time data acquisition. Machine learning (ML), a sub-discipline of AI, involves the use of statistical tools to extract the desired information either through understanding the underlying patterns in the information or by development of mathematical relationships among the critical process parameters (CPPs) and critical quality attributes (CQAs) of biopharmaceuticals. ML is still in its infancy for

Data‐driven prediction models for forecasting multistep ahead profiles of mammalian cell culture toward bioprocess digital twins

“…In addition, a hybrid approach could use an ensemble of multiple DDMs, each with different strengths and weaknesses, and combine their predictions in a hybrid manner to improve overall performance or to handle missing data points in time‐dependent parameters. Recently, linear and nonlinear DDMs were combined to develop a more accurate and comprehensive model, covering different ranges of data or features (Puranik et al, 2022). Therefore, efforts to advance data‐driven modeling will also contribute to the development of more robust hybrid models, and expand the scope of data‐driven modeling in mammalian cell culture processes.…”

“…To meet the increasing demand, the goal of biopharmaceutical manufacturing process is to quickly produce as many high‐quality biologics as possible. With the advances in process analytical technology (PAT) in conjunction with the quality‐by‐design concept (Gerzon et al, 2022), bioprocess can be promptly monitored and adaptively controlled with critical process parameters and key performance indicators, thus ensuring consistent and good quality of mAbs (Puranik et al, 2022). While a large amount of data is generated, only a limited amount of data is manageable due to the complexity and dynamics in the cellular behaviors and cultural processes as well as the intercorrelation among biological and process parameters (Gangadharan et al, 2021).…”

Can Metabolomic Approaches Become a Tool for Improving Early Plant Disease Detection and Diagnosis with Modern Remote Sensing Methods? A Review

“…The commonest types of non-parametric classifiers used for diseased and healthy plant determination are principal component analysis (PCA), support vector machine (SVM), cluster analysis (CA), partial least-square (PLS), and artificial neural network (ANN) [ 155 , 156 ]. When processing chromatographic and spectrometric data, in addition to non-parametric classifiers, databases are also used to determine recognizable substances and compounds [ 157 , 158 , 159 ]. The choice of an analysis algorithm depends on many factors, such as data amount, the presence of a visible feature’s ability to be distinguished, and so on [ 160 ].…”

Explainable AI for CHO cell culture media optimization and prediction of critical quality attribute

“…The last few years have witnessed an increasing application of machine learning (ML) approaches to deal with the amount and intrinsic complexity of biological data (Puranik et al 2022 ; Yang et al 2023 ; Rathore et al 2023 ). The typical process involves handling input data, training the fundamental model, and making predictions.…”