Reduced scale model qualification of 5-L and 250-ml bioreactors using multivariant visualization and Bayesian...

Area of Science:

• Biopharmaceutical manufacturing
• Process analytical technology
• Statistical modeling

Background:

• Reduced scale models (RSMs) are crucial for biopharmaceutical process development.
• Ensuring model equivalence across scales is vital for successful technology transfer.
• Current qualification methods may not fully capture multivariate process dynamics.

Purpose of the Study:

• To present a novel statistical method for qualifying reduced scale models (RSMs).
• To demonstrate the application of this method using data from advanced microscale bioreactors and pilot scale bioreactors.
• To ensure process performance and product quality attributes are comparable across manufacturing scales.

Main Methods:

• Utilized data from a 250-ml advanced microscale bioreactor (ambr) and a 5-L bioreactor RSM.
• Employed multivariate dimension reduction and data visualization via partial least squares discriminant analysis (PLS-DA).
• Applied Bayesian multivariate linear modeling for inferential analysis and probability distributions.

Main Results:

• The method successfully identified key process performance and product quality attributes.
• PLS-DA highlighted attributes with the greatest negative impact on multivariate Bayesian joint probabilities.
• Bayesian modeling provided probability distributions for attribute differences between scales.

Conclusions:

• The novel statistical method provides a robust framework for RSM qualification.
• This approach enables better understanding and control of scale-dependent process variations.
• It facilitates process optimization and reassessment of equivalence for improved biomanufacturing.