Tuesday March 10Tuesday, 10 March 2026 - PT (Pacific Time, GMT-08:00)

Mammalian cell culture processes have been used for decades to produce therapeutic antibodies. These bioprocesses consist of costly reagents using cell lines with long doubling times and require aseptic conditions that often necessitate expensive single use equipment. The development of an Upstream process for commercial production also involves generation of numerous data sets to demonstrate process understanding and control. The result of these numerous challenges and requirements is that the development of a process takes 10-15 years with a cost of at least $100M. The reduction of timeline and cost has the potential to be achieved through use of computational tools that have matured greatly over the last several years. Machine learning based models of the cell culture production bioreactor have been developed with the goal of not only reducing the development load but also aiding in process prediction. Examples that highlight the potential of these tools are presented.

Lab robotics are currently impaired from realizing their true potential to be zero-touch, high-throughput, tireless machines. Automation specialists, software developers, and scientists must work together for months to stand-up workflows that still need constant monitoring. At Genentech, we are working on this digital transformation challenge by experimenting with LLMs to translate 'traditional' protocols into machine-executable instructions and AI agents to orchestrate complex workflows. In this talk, I'd like to share our progress and ideas for how lab robotics can be unleashed to democratize automated experiments, ensure better error recovery, and execute workflows across different labs and instruments on demand.

• Running multiple cycles on smaller chromatography columns in dual column mode, as well as connecting multiple unit operations together in continuous fashion is a strategy to enable processing of the additional mass entering the downstream process from intensified fed batch bioreactors.
• Continuous multi-cycle operation ensures that the processing time meets or even exceeds that of standard batch methods, enabling facilities to meet their required run rates without compromising purity or yield.
• The reduction in chromatography resin volume and buffer usage enables a reduced facility footprint and lower cost of goods of manufacturing, while also improving sustainability initiatives

• Continuous downstream processing based on capture via target precipitation and polish via flowthrough chromatography offers a low cost-of-goods, low process mass index, and low complexity pathway to monoclonal antibody manufacture.
• Yield, purity, and throughput metrics for the precipitation-based process compare favorably to those for platform Protein A downstream processes.
• Capture via precipitation is broadly applicable to monoclonal antibodies and antibody-like therapeutics.
• The process development workflow in terms of selection of precipitant conditions and polishing resins is straightforward.

• When and how to plan for commercialization of your product
• Modeling supply needs and COGS of your product
• Key supply chain consideration to study early in development to improve patient access

A forward-looking discussion on how manufacturing innovation, digitalization, and improved supply chains can expand patient access and to viable business cases.

This panel will move beyond the “which is better” debate to address how leading companies evaluate and select the right upstream strategy across their portfolios. Experts will share case studies and business studies on decision-making frameworks, considering factors such as molecule type, clinical phase, cost of goods, regulatory expectations, facility readiness, and long-term scalability. Discussions will also explore the role of modeling, digital twins, and AI-driven analytics in simulating productivity and risk trade-offs between fed-batch, intensified, and fully continuous processes.

• Digital Twin integrates data, models, and process understanding from development through GMP manufacturing.
• Transitioning from post-run analysis to live predictive monitoring and control closes the loop between insight and action.
• Robust data integrity, model validation, and lifecycle management enable regulatory readiness and trust.
• Building model literacy and cross-functional collaboration across R&D, MSAT, QA, and Operations to ensure sustainable technology adoption and culture alignment.

• Develop risk-based and science-based analytical control strategy
• Considerations for assay development, qualification and validation
• Considerations for stability and shelf-life studies
• Common pitfalls & how to avoid them
  
  
  

A comparative panel on the unique manufacturing, supply chain, and scalability challenges of autologous vs. allogeneic approaches, with emphasis on future trends and digital tools for decision-making.