How our platform works
The foundation of our platform is StructureGPT, a generative pretrained transformer designed to solve the complex task of translating protein structure into its corresponding amino acid sequence. Its two key capabilities—generativity and inpainting—allow us to provide a wide range of practical solutions.
- Generativity allows us to create variants of a given protein with improved properties (e.g., stability, solubility) without altering its core function.
- Inpainting enables us to repair missing structural information or redesign sections of a protein to achieve entirely new functions, which is crucial for more innovative applications.
Complementing StructureGPT is our XSeq system, which predicts various properties of the protein sequences generated. XSeq models range from deep learning systems to more traditional AI algorithms, ensuring robust, accurate predictions. For example, our ThermoSeq model predicts a protein’s thermal stability (ΔTm) based on its amino acid sequence, helping us filter sequences to focus on those most likely to succeed in experimental testing.
Applications of our AI platform
Our platform is versatile, addressing different needs within the biopharmaceutical industry:
Protein optimization for production and manufacturing: By using the generativity of StructureGPT, we improve protein properties such as stability, solubility, aggregation, and expressibility. This service is tailored for CDMOs and biopharma companies looking to streamline production processes and reduce costs.
Development of innovative, patentable proteins: Using a combination of generativity and inpainting, we design entirely new proteins with the potential to be patented. This service is ideal for pharmaceutical R&D teams and research centers aiming to push the boundaries of biotechnology with novel therapeutics.
In addition to these primary applications, our platform has the potential to solve far more complex problems. For instance, using inpainting, we can design proteins entirely from scratch to extend the metabolism of microorganisms, enabling them to produce biofuels such as hydrogen or hydrocarbons sustainably. Similarly, our platform can be used to modify T-cell receptors to better recognize and attack cancer cells, or even to develop biosensors for detecting contaminants in water supplies.
We are still in the experimental phase, but our technology opens the door to exciting new possibilities across biotechnology, therapeutics, and industrial processes.
