Welcome to IFPAC Digital 

Bioprocess 4.0 – Vision of Sartorius

Monday, March 1 | 7:45 am – 12:00 am (EST)

Speaker:
Ulrike Lemke, Head of Marketing Bioprocess Solutions 

Bringing Bioprocess Digital Twins to Life

Monday, March 1 | 1:00 pm – 5:30 pm (EST)

Speaker:
Tiffany McLeod, Live Science Market Manager
Chris McCready, Principle Research Scientist

Abstract:
Bioprocess digital twins will change the way we do bioprocess development, validation, and cGMP operations. However, in order to bring bioprocess digital twins to fruition, a combination of IoT connectivity, statistical, and mechanistic analytics will be all required, preferably in a single environment. In this presentation, Sartorius will present how they plan to bring bioprocess digital twins to life with the new cloud based Umetrics^® Studio platform as well as the ongoing transformational initiatives across the bioprocess industry and the benefits achieved so far. These benefits include reducing time-to-market, COGS, and improving process agility and risk mitigation.

Integration of Raman Into Modern Bioprocess

Tuesday, March 2 | 12:15 pm – 12:45 pm (EST)

Room:
Sartorius Social Networking Lounge, Limited to 30 Attendees

Speakers:
Mark S. Kemper, Director of Applications Development and Customer Success, Tornado Spectral Systems, Inc.
Henry Weichert, Technology Consultant PAT, Sartorius

Sharon Deram, Director of North American Sales, Tornado Spectral Systems, Inc.

Obtaining MSPC Charts for Large Scale Based on Small Scale Studies

Tuesday, March 2 | 1:00 pm – 5:30 pm  (EST)

Speaker:
Timo Schmidberger, Senior Data Scientist 

Abstract: 
MSPC charts are important tools for effective monitoring of processes and have great value as early fault detection tools. They can be extremely valuable when a process is transferred for the first time into a larger scale in order to spot deviations as the processes evolve during transfer activities. The problem is, that typically limited or no data is available to generate powerful multivariate process models.

In this presentation, we start with a DOE and optimize a process with respect to cumulative integrated viable cell density and harvest viability. In parallel, a multivariate process model was generated as a basis for defining MSPC in a scale model.

Based on that optimization, standard operating ranges were defined and used to predict the MSPC limits. These limits were transferred into the multivariate process control model and compared to original limits and the run of a large scale run.

Within the presented approach we show a way how to use small-scale data to determine multivariate process models and the corresponding multivariate statistical process control charts for an optimized process. The generated models and MSPC limits are ready to be used with the first large scale. In this way, the first guidance of expected process behavior at scale can be obtained even before performing the first large-scale run. This can help saving development time and de-risking first large-scale runs, e.g. for the supply of clinical material.

Hybrid Modeling for Deeper Bioprocess Insight

Wednesday, March 3 | 1:00 pm – 5:30 pm (EST)

Speaker:
Catalina Moreno, Data Scientist 

Abstract: 

In laying the foundation for digital design and advanced analytics and control in next generation biomanufacturing, predictive monitoring through process modeling offers unique advantages such as process improvement and optimization.

Benefits of data-driven models such as simplicity and usefulness for real-time applications can be enhanced through a hybrid combination with mechanistic state space modeling techniques, providing greater insight to bioreactors and cell health.

In this presentation hybrid modeling will be demonstrated for an upstream bioprocess. The utility of such models in both R&D and manufacturing will be discussed, with examples including soft sensor development and digital twin simulation.

Finally, the evolution and development of commercial software addressing industry digitization will be reviewed, offering insights into the vision and roadmap of Sartorius.

Comparison of Multivariate Raman Models From Different Scales

Thursday, March 4 | 12:15 pm – 12:45 pm (EST)

Room:
Sartorius Social Networking Lounge, Limited to 30 Attendees

Speakers:
Maryann Cuellar, Life Science Product Manager Marketing, Kaiser Optical Systems, Inc.
Henry Weichert, Technology Consultant PAT, Sartorius

SIMCA Modeling for Ambr^® Cell Culture Development

Thursday, March 4 | 1:00 pm – 5:30 pm (EST)

Speaker:
Erik Johansson, Principal Data Scientist

Abstract:
Advanced data analytics facilitates a better understanding of cell culture development processes and allows root cause analysis of problematic batches. High throughput experiments such as those performed in Ambr^® 15 and Ambr^® 250 multi-parallel bioreactor systems generate large amounts of data making the data analysis particularly challenging. With SIMCA modelling an overview of all the batches from the Ambr^® runs, can be generated using a PCA model, pinpointing root cause of deviating batches, when cultures have underperformed. For the “well behaving” Ambr^® cultures, we can then visualize groups or trends among the runs and relate the information back to the measured variables. In addition, we will use OPLS to predict CQAs or a process output, in this case final titer. SIMCA can handle continuous process parameter data, data generated by integrated or at-line analysers or all data together. This presentation will focus on data from an Ambr^® 250 HT experiment and will outline the value of the SIMCA modelling. Attend and learn how advanced data analytics help in uncovering the golden nuggets among process parameters and CQAs obtained in Ambr^® cell culture development.

Advanced Monitoring and Automated Control of a Single-Use, Pilot-Scale Bioreactor via Multiple Direct Sensor Technologies With High Data Density

Friday, March 5 | 7:45 am – 12:00 am (EST)

Speakers:
Paul Reynolds, AbbVie 
Dan Kopek, Manager of Cell Culture Technologies – North America

Abstract: 
Automation frees up operators to focus on higher-level operations, and this can help ensure process robustness and success.  As we reduce the need for manual intervention for these mundane tasks, we will realize a process is faster, more granular, and closed.  We will reduce costs and produce a higher quality product in a more risk-averse and tightly controlled manufacturing environment.    

Bioreactor automation is enabled by the right sensor technologies, control strategies, and robust integration. This is especially true for single-use unit operations, where the technologies must be well-proven, not introduce additional risk, and must meet key technology attributes to achieve the desired performance goals.

This talk will outline the use of three innovative sensor technologies working in parallel on a single-use, pilot-scale bioreactor.  It will include detail on adaptation, implementation, and execution.