Automatic real‐time calibration, assessment, and maintenance of generic Raman models for online monitoring of cell culture processes

Abstract

Raman spectroscopy is a multipurpose analytical technology that has found great utility in real‐time monitoring and control of critical performance parameters of cell culture processes. As a process analytical technology (PAT) tool, the performance of Raman spectroscopy relies on chemometric models that correlate Raman signals to the parameters of interest. The current calibration techniques yield highly specific models that are reliable only on the operating conditions they are calibrated in. Furthermore, once models are calibrated, it is typical for the model performance to degrade over time due to various recipe changes, raw material variability, and process drifts. Maintaining the performance of industrial Raman models is further complicated due to the lack of a systematic approach to assessing the performance of Raman models. In this article, we propose a real‐time just‐in‐time learning (RT‐JITL) framework for automatic calibration, assessment, and maintenance of industrial Raman models. Unlike traditional models, RT‐JITL calibrates generic models that can be reliably deployed in cell culture experiments involving different modalities, cell lines, media compositions, and operating conditions. RT‐JITL is a first fully integrated and fully autonomous platform offering a self‐learning approach for calibrating and maintaining industrial Raman models. The efficacy of RT‐JITL is demonstrated on experimental studies involving real‐time predictions of various cell culture performance parameters, such as metabolite concentrations, viability, and viable cell density. RT‐JITL framework introduces a paradigm shift in the way industrial Raman models are calibrated, assessed, and maintained, which to the best of authors' knowledge, have not been done before.

Technology outlook for real‐time quality attribute and process parameter monitoring in biopharmaceutical development—A review

“…Real‐time machine learning tools for the prediction of product concentrations in mammalian cell culture were reported by Bayrak, Wang, Tulsyan, Coufal, and Ündey (2018), where quality attributes were forecasted in advance for better control of the process. A just‐in‐time machine learning framework for automatic calibration, assessment, and maintenance of generic Raman models for on‐line monitoring of cell culture processes was recently reported (Tulsyan, Wang, et al, 2019). This approach of a just‐in‐time generation of Raman models for on‐line monitoring is a unique alternative to the traditional method where models are created based on a historical training data set.…”

“…Upon identification of any deviations of model‐predictions from reference assay values or historical data trends, model amendments can be implemented such as extensions, corrections, advanced‐revisions or re‐development of the model. Tulsyan and coworkers recently reported a machine learning approach to automatically conduct model maintenance in real time for generic Raman models for glucose, glutamine, VCD, and lactate in cell culture process (Tulsyan, Wang et al, 2019). A detailed discussion of MLM and model maintenance strategies in biopharmaceutical applications are discussed elsewhere (Flåten, 2018).…”

Portable Analytical Techniques for Monitoring Volatile Organic Chemicals in Biomanufacturing Processes: Recent Advances and Limitations

“…Especially for off-line methods, pretreatment of culture medium such as filtration and isolation could significantly improve the detection sensitivity and specificity. Nevertheless, sampling frequency (once every 12–24 h) in off-line contexts is often too low to monitor highly dynamic metabolic processes (Tulsyan et al, 2019 ), which may result in limited spatiotemporal resolution of the cell metabolic features. Moreover, the pretreatment procedures are usually time-consuming and laborious, and the cultivation fluid is subject to contamination caused by repetitive sampling.…”

“…In addition, owing to fast data acquisition, high accuracy, and capability of multiparameter analysis, Raman spectroscopy is considered a promising tool in monitoring a wide range of aqueous biological samples (Shaw et al, 1999 ). For instance, Santos et al ( 2018 ) monitored monoclonal antibody (mAb) cultivations in situ using Raman spectroscopy by adjusting calibration models; an automatic and real-time calibration framework was established by Tulsyan et al ( 2019 ) that enabled the integration of traditional Raman models in variable culturing conditions without specific calibration steps, thus increasing the feasibility of Raman techniques in inspecting culture media with highly changeable composition.…”

Fiberoptic-Coupled Spectrofluorometer with Array Detection as a Process Analytical Chemistry Tool for Continuous Flow Monitoring of Fluoroquinolone Antibiotics

“…PAT has many applications in the pharmaceutical and antibiotics manufacturing [12][13][14][15][16][17][18][19][20][21][22][23][24][25], chemical [26,27], petrochemical [28], and food industries [29][30][31][32][33]. In addition, there are many recent progress in realtime monitoring of cultivations in bioreactors and cell culture process [34][35][36][37][38][39][40][41], fermentation [42,43] and biological process [44], and electrochemical [45,46] and protein purification [47]. Using PAT enables us to get a deeper understanding of the process.…”