Digitalization is one of the buzzwords that is on every organization's goals. The talk will discuss what digitalization is and what it is not.

In this presentation, we show how advanced machine learning and hybrid modeling approaches can be exploited to significantly improve process understanding, performance, and automated operation as digital twins. All presentations will be centered on industrial implementation examples for mAb, cell & gene therapy, and mRNA processes with numerous big pharma and CDMO partners allowing to quantify efficiency gains in process development.

Implementing a framework for unified IT/OT infrastructure and data management is foundational to digital process development. This framework of connected systems, data and analytics can be harnessed to allow end-to-end process visibility, improve comparability and predictability across scales and products, and thus enhance productivity and quality. A holistic & pragmatic strategy will be discussed to advance data architecture, governance, integration, sharing and analytics consumption in support of our core mission to bring new medicines to patients

Sanofi CMC/Process Development is implementing a comprehensive Digital Transformation program to improve our productivity and reduce our development timelines. We are deploying a standardized digital workflow across our development sites globally (target: 2,000 users) and building automated data access to ensure scientists can document, visualize, and analyze their experimental work. We will discuss the progress, successes, and pain-points encountered by the development teams as we progress towards our digital ambition.

Acceleration of commercialization of biologics including the filing of a robust control strategy is of utmost importance for biosimilars up to new modalities. Digital twins capture CMC knowledge and allow multiple deployments. We will show

• How end-to-end digital twins can help saving 50% of experimental effort by incorporating drug substance specification when designing unit operations and
  
• How real time application allows for prediction and control on process performance

• Critical inputs for creating a representative model
• Extracting actionable data and insights from simulations
• Validating results and impacting the process development outcome

• ​Process integration a prerequisite for automation
• What are the barriers for full automation?
• Which sensors do we need?
• Benefits of automation
  

The long-term vision of biomanufacturing is autonomous bioprocessing. To achieve this state, it is necessary to automate bioprocesses, which require sensors to control a manufacturing system. Currently, for a lot of quality/process parameters sensors are not available and soft sensors and self-learning algorithms must be applied. The state-of-the-art of monitoring and control of bioprocesses will be provided and to which extent integrated continuous biomanufacturing necessitates autonomous bioprocessing.

Model-based approaches in the biopharma industry have the potential to accelerate decision-making, optimize the product to market time, and reduce costs. In silico representation of manufacturing processes is becoming easier thanks to analytical tools, process modeling software, and machine learning algorithms. In this contribution, we demonstrate the use of model-based approaches for decision-making for ultrafiltration and formulation processes.

Biopharmaceutical industry traditionally relies on pharmaceutical manufacturing practices to monitor processes and release products. The use of Process Analytical Technologies (PAT) can improve the process monitoring and control and at the same time increasing the process understanding and modelling capabilities. This talk examines the possible PAT application to vaccine processes using a phase- and technology-appropriate approach to reach a fully implemented in-line monitoring (ILM).

Advanced therapy medicinal products (ATMPs) require new manufacturing technologies for industrialization. Especially allogenic cell therapies require novel approaches to move from highly manual cell cultivation processes towards automated and closed systems that allow industrial scale-up or scale-out, while ensuring sufficient process analytical technologies and interfaces for real-time monitoring and control.Here we will present our recent developments in PAT and automation for transferring manual adherent iPSC cultivation processes to highly automated and fully closed cultivation systems. New PAT methods allow for rapid in-process testing of parameters specific to these new modalities and in combination with our robotized cell cultivation platform, the transition from a laboratory process to a closed and data-driven industrialized process is enabled.

For many modern biopharmaceutical processes, manufacturers develop data-driven models using data analytics/machine learning methods. The challenge is how to select the best methods for a specific dataset to construct the most accurate and reliable model. This presentation describes the application of smart process data analytics software to industrial end-to-end biomanufacturing datasets for monoclonal antibody production to automate the determination of the best DA/ML tools for model construction and process understanding.

Raman spectroscopy has been utilized for many different applications in cell culture bioprocesses, however, the adoption of PAT into commercial environments has been slow. A toolbox methodology will be shown with the aim of reducing time to implement Raman spectroscopy applications in cell culture processes.