Increasingly elaborate and voluminous datasets are generated by the (bio)pharmaceutical industry and are a major challenge for application of PAT and QbD principles. Multivariate data analysis (MVDA) is required to delineate relevant process information from large multi-factorial and multi-collinear datasets. Here the key role of MVDA for industrial (bio)process data is discussed, with a focus on progress and limitations of MVDA as a PAT solution for biopharmaceutical cultivation processes. MVDA based models were proven useful and should be routinely implemented for bioprocesses. It is concluded that although the highest level of PAT with process control within its design space in real-time during manufacturing is not reached yet, MVDA will be central to reach this ultimate objective for cell cultivations.
- process analytical technology
- principal component analysis
- monitoring batch processes
Mercier, S. M., Diepenbroek, B., Wijffels, R. H., & Streefland, M. (2014). Multivariate PAT solutions for biopharmaceutical cultivation: current progress and limitations. Trends in Biotechnology, 32(6), 329-336. https://doi.org/10.1016/j.tibtech.2014.03.008