Deeptech drug discovery company Qubit Pharmaceuticals today unveils what it calls the world’s most advanced quantum AI model to unlock an entirely new range of therapeutics at a fraction of the time and cost of current drug discovery efforts.
According to the company, the quantum AI model, developed in partnership with Sorbonne University, is capable of modeling and simulating the behaviour of molecules with a level of precision and computational speed never before achieved.
Predicting a drug's ability to bind to a protein (or to RNA or DNA) is one of the most complex tasks in drug discovery. Chemical space is virtually limitless. An infinite number of drug molecules can be designed and introduced into an endless number of targets (around 100,000). There are thus trillions of possible combinations, and it's impossible to fit them all into a database.
The accuracy will massively reduce the costly stage of laboratory experimentation in developing new molecules by replacing the chemical synthesis of drug candidates.
“This simulation method would reduce the cost of the drug discovery phase enormously,” says Jean-Philip Piquemal, Professor at Sorbonne University and Director of the Theoretical Chemistry Laboratory (Sorbonne University/CNRS), co-founder and Scientific Director of Qubit Pharmaceuticals.
“The model is as accurate as the experiment; we can generate many new ideas, failing quickly and cheaply ‘in silico’ before moving on to laboratory testing with molecules that have passed the tests with flying colours.”
Creating foundational model FeNNix Bio1
The team used unprecedented computing power from GENCI, EuroHPC, and Argonne to create FeNNix Bio1, a foundational model built on millions of meticulous molecular simulations.
It was trained on what it calls the world's most accurate molecular chemistry database, simulated to the highest possible chemical precision. By training on these elementary bricks, the foundation model learns the laws of chemistry and physics and can reconstruct biomolecules in a Lego-like fashion. It also learns how molecules interact with each other.
FeNNix-Bio1 cracks molecular modelling’s hardest challenge: water
FeNNix-Bio1 has proved its effectiveness in one of the most difficult tasks in molecular modeling: simulating the physical behaviour of water in its various phases.
Indeed, the foundation model can accurately predict various physical properties and reproduce the behavior of ions and small organic molecules in solution with remarkable fidelity, where other reference models are unable to.
This is essential because water is the solvent present in the human body, and its interaction with drugs plays a key role in their activity.
Going beyond AlphaFold
The Sorbonne University research team that developed FeNNix-Bio1 set itself the goal of going beyond the capabilities of AlphaFold, the AI software developed by Google DeepMind, which provides a protein structure prediction based on their amino acid sequence. But FeNNix-Bio1 goes further.
According to Jean-Philip Piquemal, AlphaFold has revolutionized protein structure prediction.
“However, proteins are not static; their structures evolve over time, modifying drug interactions.
FeNNix-Bio1 makes it possible to model these dynamic effects.
In addition, AlphaFold does not accurately model the interactions of proteins with drug candidates.
FeNNix-Bio1 addresses these two important limitations for biomolecular simulation.”
Quantum-level accuracy while remaining scalable and cost-effective
Where traditional simulations are limited in accuracy, speed, and areas of applicability, and quantum chemical models are accurate but slow and too computationally expensive for large-scale implementation, FeNNix Bio1 offers quantum-level accuracy while remaining scalable and cost-effective. The foundation model doesn't just predict structure, it understands how molecules behave and interact.
To achieve this, the FeNNix-Bio1 researchers developed neural network approaches adapted to applications in chemistry and physics, rather than using LLMs architectures, which are generally optimised for recognising and generating text.
More accurate and less expensive, FeNNix-Bio1 can be trained in a few hours using a standard GPU, whereas other AI models require weeks of supercomputing time.
“We're aiming at complex targets, those for which the pharmaceutical industry doesn't provide a solution for patients,” comments Robert Marino, CEO of Qubit Pharmaceuticals.
The model can simulate any system
The company currently has a pipeline of 7 programs, particularly in oncology and inflammation. The most advanced program concerns breast cancer.
Rooted in the laws of physics, the model is adaptable: By changing the molecular building blocks, any system can be simulated.
Further potential applications in the chemical sector include the design of industrial enzymes, the optimisation of membranes for desalination, the development of new-generation batteries, and the acceleration of green chemistry.
Quantum AI in action
FeNNix-Bio1 also paves the way for quantum AI — the convergence of quantum computing and machine learning, which promises to revolutionise data generation for molecular simulations.
According to Piquemal, Qubit Pharmaceuticals is already using quantum data to enrich its models, something that was long thought impossible until 2035.
Lead image: Robert Marino, CEO of Qubit Pharmaceuticals. Photo: uncredited.
Would you like to write the first comment?
Login to post comments