This presentation will highlight our manufacturing platform strategy for cell therapies—emphasising data science, modelling, and PAT—to enhance manufacturing efficiency and product quality. By integrating automation and machine learning, we want to accelerate the development of robust processes. Join us in exploring how smart manufacturing practices can redefine the future of cell therapy production.

A novel program of work (Digital_Lyo) will be presented that is being undertaken by a consortium or academic, industrial, and regulatory authority partners, including AstraZeneca, Siemens, the Medicines and Healthcare Regulatory Agency (UK), and smaller industrial enterprises with specialist capability in sensor development. The talk will present highlights of the Digital_Lyo programme, including the applications for a novel process analytical technology called through-vial impedance spectroscopy (TVIS).

Cell culture development in biopharmaceuticals uses models to optimise processes and understand complex systems. Defining a model's purpose is crucial, as is starting with good data. Embracing model thinking across all disciplines enhances understanding and effective use of models. Despite challenges, effective modelling can lead to faster, better, and more sustainable processes.

Bio-capacitance has become a standard online method to measure growth in cell-based biomanufacturing. The method offers, rapid continuous monitoring without manual sampling. However, there are noted deviations at the inflection point beyond exponential growth compared to standard staining methods such as trypan blue. This can be explained by different measurement criteria that can be exploited to gain a good understanding of the metabolic changes that arise during the bioprocess.

Digital twins can significantly transform bioprocess design. We present a case study on AAV production in insect cells, where digital simulation was used to design a new fed-batch operation. This demonstrate how digital twins can be leveraged for new gene and cell therapy products, which often have limited accumulated data and serotype/donor-specific variability, to unlock faster and more efficient bioprocess development.

The current workflow involves experimental DOE to determine the optimal culture media formulation. A paradigm shift is underway in the optimisation of culture media, wherein a modelling approach can be employed to accelerate culture media optimisation. I will introduce a computational approach involving genome-scale metabolic modelling and model-guided DoE approach, and how this workflow can help the industry, particularly for new modalities, to accelerate culture media design and optimisation.

rAAV processes with complex dynamic behaviour requires high experimental effort, and is time consuming and expensive. Digital Twins (DT) that are based on mathematical models can be used for process development and optimisation. A mechanistic model of an rAAV9 production process has been developed and parameterised using in-house experimental data. The validation of the DT models and its application to increase the functional rAAV9 titre will be demonstrated.

Orchard’s approach to gene therapy is designed to deliver a functional version of the mutated gene, or transgene, to a patient’s own blood stem cells—called hematopoietic stem cells or HSCs—to produce the desired therapeutic protein. This talk will discuss lentiviral process development and scale-up.

Focusing on regulatory challenges, this presentation explores the role of closed processing in Advanced Therapy Medicinal Products (ATMPs). Closed systems, together with standard process platforms and automation-digitalisation, are key for the future of ATMP manufacturing. Through the analysis of the current guidelines, this presentation draws how regulatory frameworks may either support or hinder closed systems adoption. The final goal is to align closed processing with current and future regulations.