The pharmaceutical industry is on the cusp of a transformative era driven by advancements in artificial intelligence (AI) and innovative chemical processes.
An event sponsored by CPA (Chemical pharmaceutical generic association) and entitled Artificial intelligence as the future frontier for chemical and pharmaceutical innovation explored the world of artificial intelligence and its impact on the chemical and pharmaceutical world.
The conference adopted a holistic approach, grounded in the understanding that knowledge gained from studies and experience should be shared globally.
We have a task and that is that of a world order, of disseminating the knowledge we have acquired.
Marcello Fumagalli, CPA’s General manager
The sharing of knowledge and scientific discoveries can no longer be restricted to a select few or locked away in inaccessible archives.As professionals in this field, it is imperative to embrace the dissemination of knowledge, ensuring that groundbreaking research and insights do not remain confined but are shared widely.
So the conference addresses topics ranging from the first algorithm to the future of the quantum computer, from social aspects (risks and fears, impact on society) to operational ones (with case studies from the real world) to philosophical ones (will Ai be able to have a conscience of its own?).
A deep dive into the abysses of artificial intelligence in the world of chemistry and pharmaceuticals.
The integration of AI in drug design and development promises to revolutionize our approach to therapy, enabling precise targeting of pathological states with reduced investment. However, the challenge remains in navigating regulatory frameworks that are often slow to adapt to such rapid advancements. By fostering a culture of knowledge sharing and collaboration, we can overcome these hurdles and contribute to a global effort in advancing pharmaceutical research and development.
Here we explore concrete applications implemented by some Big Pharmas showcasing how AI is transforming their operations.
Product Formulation with GSK
A significant AI application in product formulation is University of Padova collaboration with GSK to optimize tablet production. Using advanced digital models, AI assists in correctly dosing lubricants, minimizing the number of required experiments. This approach not only reduces development costs but also ensures that final tablets have the desired properties for effective dissolution and absorption in the human body. The semi-empirical methodology adopted saves up to 70% of traditional experimental efforts.
Product Quality Monitoring with Sanofi
In product quality monitoring, AI plays a crucial role in developing integrated systems that control quality during production. An example is Univeristy of Padova’s project with Sanofi for avian vaccine production, where an advanced monitoring system controls critical variables like pH and temperature in real-time. Virtual sensors predict the final product quality halfway through the production cycle, allowing timely corrective actions and significantly reducing waste.
3D Lung Reconstruction
Another innovative AI application at Chiesi involves the 3D reconstruction of patients’ airways. Using specialized neural networks, medical scans are processed to create accurate 3D models of the respiratory system. This enhances the understanding of drug deposition in the lungs and supports the development of more precise pharmacokinetic and pharmacodynamic models. These models optimize dosages for clinical studies and aid in designing more effective inhalation devices.
Molecule Prioritization in Pre-Clinical Phases
In pre-clinical stages, Chiesi has integrated AI to enhance the prioritization of candidate molecules. Traditionally reliant on statistical approaches, the introduction of graph neural networks has revolutionized this process. These networks predict critical physicochemical properties of molecules, such as lipophilicity, solubility, and stability. This allows for the selection of the most promising molecules for further synthesis and characterization, optimizing time and resources invested in developing new active ingredients.
Hybrid monitoring systems
Furthermore, hybrid monitoring systems are innovating by combining real data with digital models for more accurate and timely diagnostics of production processes. These systems, such as those used in powder drying processes, utilize digital twins to create virtual variables that enhance the ability to anticipate malfunctions and diagnose issues, ensuring greater efficiency and reliability in pharmaceutical production. Additionally, machine learning is essential in improving the transfer and scale-up of production processes, like monoclonal antibody production, by analyzing metabolomic data to optimize cell line responses and feeding strategies, leading to more efficient production and higher quality products.
