Mini‐review: soft sensors as means for PAT in the manufacture of bio‐therapeutics

Abstract

This mini‐review discusses how soft sensors can contribute to accomplish FDA's process analytical technology (PAT) ambitions in the manufacture of bio‐therapeutics. Focus is on applications with protein‐based drugs (proteins, antibodies), but also gene therapy vectors as well as cell cultures are considered where chemical and bio‐analytical as well as mathematical and statistical methods are used as tools. An overview of existing soft sensor alternatives and how these can be configured to meet typical industrial needs is provided. It is noted how several of these needs coincide with the PAT regulatory incentives but do also address process economic aspects of bio‐therapeutic manufacture. Evaluation of soft sensor alternatives is highlighted in relation to the production targets, quality attributes and the specification of these as well as shortcomings and needs for further improvements. © 2014 Society of Chemical Industry

Control of specific carbon dioxide production in a fed-batch culture producing recombinant protein using a soft sensor

“…Moreover, the soft sensor method further underscores the PAT mind-set of applying on-line sensor-based control in biotherapeutics manufacturing (Glassey et al, 2011;Mandenius and Gustavsson, 2015) with focus on reducing product variability. However, the method is applicable to most aerobic fed-batch cultures with recombinant proteins provided the CO 2 production is high enough.…”

Process‐wide control and automation of an integrated continuous manufacturing platform for antibodies

“…Alternatively, atline analytical devices, such as HPLC systems, can give important process insight along the sequence of unit operations in the manufacturing platform (Karst et al, 2017). In cases where CQAs cannot be directly assessed, soft‐sensor‐based technologies can provide relevant substitutes (Kroll, Hofer, Ulonska, Kager, & Herwig, 2017; Mandenius & Gustavsson, 2015; Narayanan et al, 2019; Sokolov, Feidl, Morbidelli, & Butte, 2018; Solle et al, 2017; Sommeregger et al, 2017).…”

Decision Tree‐PLS (DT‐PLS) algorithm for the development of process: Specific local prediction models

“…For the broad and automated integration of this technology into the R&D and manufacturing ecosystem, data‐ and knowledge‐driven model‐based solutions play a key role. Data‐driven methods are centered on the generated process data sets and target enhanced process understanding and prediction through visualization, correlation analysis, and regression techniques . Given the complexity and variety of bioprocess data, a customization of data‐driven tools on the basis of process specificities and analysis expertise is crucial for their consistent utilization.…”