SC5: Saving Time in Process Development with Next-Generation Methods: iDoE, Hybrid Modeling and PAT

 

  

18:30     QbD, PAT and different types of process models

Agenda of Part 1:  

What are the advantages of Quality by Design (QbD) over Quality by testing? How can Process Analytical Technology (PAT) and process modeling support to speed up bioprocess development while meanwhile reducing batch to batch variations and out of specification events? What are the differences between parametric (white box or mechanistic), non-parametric (black box or statistical) and semi-parametric (hybrid) models and where can we apply those methods? For which process operations should I use statistical, mechanistic or hybrid models?

19:20 Dinner

19:45    Introduction to the Hybrid Modeling Toolbox 

In the second part, hybrid modeling and the concept of intensified Design of Experiments (iDoE) for process optimization and characterization will be discussed and their application for a cell cultivation process will serve as example. Based on the previous discussed advantages of hybrid models a hybrid modeling toolbox will be presented.  The handling of the toolbox and the benefits will be demonstrated on both a DoE and intensified Design of Experiments (iDoE) E. coli fed-batch data set. It will be shown how the characterization times can be reduced by over 40% using this unique approach. 


Agenda of Part 2:

  • Introduction to the toolbox (display, data preprocessing, workflow)
  • Concept of iDoE
  • Setup of Hybrid Models
  • Demonstration of the ability to reduce process development and characterization times by applying the toolbox on iDoE data
  • Participants can bring their own laptops (windows only) and directly follow the workflow. A free version of the toolbox is available here.

Learnings from this Short Course:

  • Why process modeling is a key to boost bioprocess understanding and development!
  • Hurdles to overcome before you start modeling!
  • Awareness of different type of process models!
  • Usage of those different models for different unit operations!
  • Handling of a use-case centralized hybrid modeling toolbox!

21:00     End of Course

Instructor Biography

Duerkop_MarkMark Dürkop, PhD, CEO, Novasign GmbH, and Principal investigator, University of Natural Resources and Life Sciences, Vienna (BOKU)

With more than ten years of experience, Mark is a passionate scientist in the field of biopharmaceutical process development. He worked together with several Top25 Pharma companies in different projects targeting their needs in process development and understanding. Together with his team, he is developing a hybrid modeling software for both up and downstream processes. This software utilizes AI applied to process data and first principle models (hybrid models) to quickly predict critical quality attributes (CQAs) from critical process parameters (CPPs). Hereby, bioprocess development and characterization is accelerated significantly while simultaneously a better process understanding is gathered. From this research project, the company Novasign was founded offering this unique software solution.