Industry 4.0 with Lonza
Lonza has had a busy opening to 2024, most notably the company announced that it has had a $1.2b offer accepted by Roche Pharmaceuticals to acquire the Genentech large-scale biologics manufacturing site in Vacaville, California. This acquisition aims to significantly increase Lonza’s large-scale biologics manufacturing capacity to meet demand for commercial mammalian contract manufacturing from customers with existing commercial products, and molecules currently on the path to commercialisation within the Lonza network.
Lonza plans to invest approximately CHF 500 million in additional CAPEX to upgrade the Vacaville facility and enhance capabilities to satisfy demand for the next generation of mammalian biologics therapies. The products currently manufactured at the site by Roche will be supplied by Lonza, with committed volumes over the medium term, phasing out over time as the site transitions to serve alternative customers.
Over the final half of 2023, we caught up with a vast range of Lonza SMEs discussing everything from the process and benefits of spray-dried dispersion technologies to the key targets of Lonza’s Micronisation Centre of Excellence. This year, we plan to stay in the Lonza loop once again by continuing to speak with some of their internal leaders, pushing the industry ever-further forward.
In our first interview series of the year with Lonza, we spoke with Dr. Simon Wagschal, Associate Director, Advanced Chemistry Technologies, Lonza Small Molecules Global R&D and Dr. Aaron Johnson, Manager, Cheminformatics & Data Science, Lonza Small Molecules Global R&D.
Jai: For Lonza, we know the company aims to remain at the forefront of innovation. In what ways has the company been able to capitalise on Machine Learning and what have been the stand-out developments thus far? (long answer if possible – detailing a few developments would be ideal, please)
Dr. Simon Wagschal: The biotech and pharmaceutical industry continues to evolve and Lonza is harnessing innovations in Machine Learning (ML) to meet our customers’ needs and help bring their innovations to market. At Lonza, ML is used in synthetic process route design, process optimisation, toxicological assessment of new chemical entities, and formulation design.
While all the above-mentioned applications are driving efficiencies in our ways of working, we have recently appreciated ML-driven gains in our Early Phase service offerings. Small molecule API process route scouting is one example. In this aspect of IND-enabling work, process chemistry SMEs devise prospective synthetic pathways from commercially accessible starting materials, through successive intermediates of increasing molecular complexity, to arrive at a target active pharmaceutical ingredient. After prioritising the possible routes for alignment with clinical and commercial manufacturing ideals, the team experimentally assesses the performance of the top-rated options. The advent of computer-aided synthesis planning technologies (CSPTs) with route prediction capabilities has materially impacted this workflow. Virtually all CSPTs use a subset of machine learning, called Monte Carlo Tree Search, to quickly prioritise computationally predicted synthetic routes. This heuristic search approach, in our context, prioritises prospective routes by iteratively computing probabilities of reaction success at each step in a proposed pathway between starting materials, intermediates, and target API. The resulting prioritised routes are served to the SMEs for consideration and further adaptation. Guided by our process chemistry SMEs, this has spurred larger volumes of prospective strategies and a higher incidence of shorter API syntheses.
Jai: What role does artificial intelligence (AI) play regarding route design, cost-effective modifications, and overall timelines for route scouting?
Dr. Aaron Johnson: Lonza is also using AI to help address the growing need to develop therapies faster and more efficiently – ultimately reducing time and cost for customers. One area where AI is being leveraged is the manufacturing of active pharmaceutical ingredients (APIs).
APIs are continuing to grow more complex, impacting speed-to-clinic. These longer synthetic pathways present challenges for process chemists hoping to achieve an efficient API manufacturing process. Overcoming these challenges is crucial to mitigate delays and expenses associated with bringing new drug candidates from early development to commercial production.
To address these obstacles, we are leveraging predictive route design technology integrated with a highly curated and expansive supply chain database. This combination yields time and cost savings for developing complex APIs. As the industry encounters more drug candidates requiring over 20 synthetic steps on average, AI-powered predictive route design technologies have become indispensable for designing efficient retrosynthetic routes. However, these tools have primarily been designed for early phase discovery phases, often producing retrosynthetic routes not optimised for commercial-scale production realities. To bridge this gap, we have merged leading AI-powered predictive route design models with proprietary supply chain data spanning raw material costs, availability, and comprehensive supply chain intelligence. This integrated technology empowers process chemists to rapidly pinpoint not merely the shortest pathways, but those most commercially viable by factoring in real-world constraints. By providing this holistic view, these cutting-edge AI solutions streamline route scouting and ultimately accelerate the transition from research and development to robust manufacturing processes.
Jai: The prominence of “industry 4.0” has been immensely significant on the pharma industry, yet there are undoubtedly challenges that arise when implementing new technologies to enhance drug development. What challenges have Lonza faced and how have they been overcome?
Both: The prominence of Industry 4.0 has brought significant challenges for CDMOs like Lonza in implementing advanced technologies like AI to enhance drug development at a commercial scale. Some key challenges Lonza has faced include:
1. Adapting and modifying AI systems to provide meaningful insights for large-scale manufacturing operations.
2. Integrating and managing diverse data sources effectively.
3. Upskilling the workforce to leverage AI technologies effectively.
To overcome these challenges, Lonza has taken a strategic approach, including forging partnerships and collaborations, implementing AI through pilot projects and iterative rollouts, and providing comprehensive employee training. By addressing these hurdles, Lonza has successfully integrated AI and other Industry 4.0 technologies into its operations, leading to more efficient and data-driven decision-making, streamlined processes, and enhanced quality control in drug development and manufacturing.