As biopharmaceutical programs advance under increasingly compressed timelines, upstream development teams face pressure to deliver optimized, quality-targeted processes earlier. Early decisions still rely heavily on platform knowledge and expert judgment, while historical bioreactor data is rarely leveraged quantitatively. This webinar will explore how hybrid modeling can bring predictive, model-based decision-making into upstream development earlier, helping optimize processes during clone selection.
Central to this approach is knowledge transfer: reusing data from prior programs to build a reliable hybrid model when new-program data is still scarce. The session will examine how relevant historical programs are identified using a multivariate strategy that scores candidate datasets on product-quality similarity, process-performance similarity and design-space coverage. These models can then support
in-silico
, clone-specific optimization. Because the model is established early, its predictions can be confirmed within experiments already planned in the workflow, reducing the need for dedicated optimization studies.
Through two case studies, the featured speakers will discuss how the same strategy adapts to different programs and quality objectives. The first, a monoclonal-antibody intermediate of an antibody-drug conjugate (ADC) developed under a compressed resupply timeline, used model-guided optimization to reduce acidic variants while maintaining titer without additional experiments. The second, another ADC intermediate at first-in-human development, applied the same approach to optimize glycan quality attributes. Together, these examples demonstrate how model-based optimization can control multiple product-quality attributes earlier in development.
Attendees will understand how integrating historical data, hybrid models and a small amount of molecule-specific experimentation can replace sequential, intuition-driven optimization with model-based decision-making. This approach reduced development timelines by up to six months and experimental runs by up to 50% relative to conventional workflows.
Register for this webinar to learn how hybrid modeling can accelerate upstream process development, reduce experimental requirements and support earlier control of product-quality attributes.
Who Should Attend
This webinar will appeal to:
Process Development Scientists
Cell Line Development Scientists
Process Development Leads / Directors
Data Scientists in BioPharma R&D
What You Will Learn
Attendees will gain insight into:
How historical bioreactor data can support hybrid modeling when molecule-specific data are limited
How multivariate dataset selection can identify relevant prior programs based on product quality, process performance and design-space coverage
How hybrid models enable
in-silico
, clone-specific process optimization earlier in upstream development
How model-first optimization reduced development timelines by up to six months and experimental runs by up to 50%
Register free on Xtalks:
https://xtalks.com/webinars/accelerate-upstream-process-development-using-hybrid-modeling/
