Discovering an effective and safe therapeutic is only the first step in the pharmaceutical development process. You need to deliver it in a formulation that meets practical constraints while ensuring the right dosage is achieved, safely and cost-effectively. And you need to manufacture the product at scale, which can be particularly challenging for biopharmaceuticals.
Alchemite applies powerful machine learning to guide your testing program and identify formulation and process improvements that deliver better product performance and control costs, while meeting regulatory constraints. It gets more value from your existing data, even where that data is sparse and noisy, and helps you to find the most efficient routes to improve the quality of this data and your understanding of it.
Example projects
New project targets oligonucleotide manufacturing
Intellegens and the Centre for Process Innovation (CPI) have secured a £1.6m grant from Innovate UK, part of UK Research and Innovation (UKRI) to apply machine learning (ML) in the emerging field of oligonucleotide therapies. This class of therapeutics holds great promise, but progress is constrained by limits on industry-wide manufacturing knowledge. The project is applying ML to identify productive routes in oligonucleotide process development and manufacturing, drawing on expert support from six leading pharmaceutical research organisations, including top 10 pharma companies such as Novartis and specialist biotech companies such as Silence Therapeutics.
Recorded webinar – biopharmaceutical case study
This webinar included a case study from Lukas Kuerten of CPI on using machine learning to predict the manufacturing viability of a class of biopharmaceuticals, monoclonal antibodies (mAbs). These molecules are too large for synthesis and must be grown organically in cells. This is a challenging research problem as there is a need to deliver predictable outputs from a living system within a tight regulatory framework. Processes and raw materials are high cost and often very low volumes, resulting in sparse datasets. Alchemite™ machine learning has been applied to extract valuable information from this data, aiming to focus development work on the most promising candidates.
Food formulation development at Yili
Pharmaceutical and biotech companies can learn from the experience of other industries. Global dairy producer, Yili, gained valuable insights into the behaviour of food formulations over time, including understanding that enabled them to accelerate their development of improved recipes.
“We relatively quickly could drop out a number of the ingredients we had been testing… This wasn’t obvious if you just looked at them one-by-one, because you always have some cross-interactions. This was a big learning and helped speed up the development.”
– Matthias Eisner, Innovation Manager, Yili
Alchemite™ for formulations and manufacturing
Alchemite™ is ideal for quick response to formulation or manufacturing challenges. Key features are:
- Gap-fill and validate sparse, noisy data from suppliers, experiment, simulation, and production
- Auto-generate models that identify key ingredient-process-property relationships
- Quantify uncertainty to support a rational business case for key decisions
- Design experimental programs, reducing the number of experiments by up to 90%
- Identify new or improved formulations and optimise process parameters
- Capture knowledge of formulation recipes as models for sharing and re-use
