ff
Artificial Intelligence & Machine Learning for Accelerated Crystal Engineering
Towards Future of Crystal Engineering
In the MacGillivray Group, a core pillar of our research is advancing Crystal Engineering for Accelerated Material Design. To push the boundaries of this field, our team is harnessing the power of Artificial Intelligence (AI) and Machine Learning (ML) to predict the formation of new solid-state materials, with a specialized focus on cocrystals.
Traditionally, the discovery of novel materials has relied heavily on resource-intensive, trial-and-error laboratory experiments. Our project aims to completely transform this workflow. By leveraging vast amounts of historical structural data, we are training advanced AI models to understand the complex rules that govern how different molecules interact, fit together, and crystallize in three-dimensional space.
Project Goals and Applications
The primary goal of our AI initiative is to rapidly and accurately evaluate millions of potential molecular combinations in a virtual environment. By pinpointing the most promising chemical candidates before any physical experiments are conducted, we can save significant time, reduce research costs, and minimize chemical waste.
Our predictive models are being developed to target several high-impact areas of crystal engineering:
- Pharmaceutical Cocrystals: Rapidly identifying safe coformers that can combine with Active Pharmaceutical Ingredients (APIs) to enhance critical properties like drug solubility, stability, and bioavailability.
- Targeted Synthons: Engineering materials by predicting specific intermolecular interactions (such as tailored hydrogen or halogen bonding networks) to create highly reactive or functional solids.
- Specific Geometrical Properties: Designing cocrystals that meet exact spatial, volumetric, and structural requirements for advanced applications in green chemistry and organic materials.
Multidisciplinary Leadership & Collaboration
Moving the needle in AI-driven chemistry is an inherently multidisciplinary challenge that bridges laboratory synthesis with advanced computer science. This theme of research in our group is led by Dr. Farshid Effaty (Institut Courtois and FRQ Postdoctoral Fellow) in collaboration with Professor Bang Liu (Canada CIFAR AI Chair at Mila), both of whom are affiliated with the Institut Courtois and MILA – Quebec AI Institute.
Their combined expertise in deep learning, artificial intelligence architectures, and machine learning is translating theoretical crystal engineering principles into robust, predictive computational tools. Together, we are moving the field away from experimental guesswork and paving the way for a faster, greener, and more rational era of materials design.