·
·
Tech & Science
Researchers used 94 qubits to model proteins with over 12,000 atoms, marking a 40-fold scale-up in quantum chemistry simulations.
Scientists from Cleveland Clinic, RIKEN, and IBM have achieved the largest quantum-classical chemistry simulation ever performed, modeling protein-ligand systems containing more than 12,000 atoms. The breakthrough demonstrates how quantum computers can work alongside classical supercomputers to tackle real-world chemistry problems at an unprecedented scale.
The team simulated two biologically important proteins, T4-Lysozyme and Trypsin, along with their binding molecules, all within a realistic water environment. The largest system comprised 12,635 atoms and approximately 30,000 orbitals, far exceeding previous quantum computing demonstrations in chemistry.
This achievement comes just months after researchers modeled a much smaller 303-atom protein. The new work represents a 40-fold increase in system size and a 210-fold improvement in accuracy for a critical part of the workflow, underscoring the rapid pace of progress in the field.
To accomplish this, the researchers combined quantum processors with high-performance classical systems in what they describe as a quantum-centric supercomputing workflow. Quantum hardware handled the most computationally demanding parts of the calculation, while classical supercomputers assembled the results.
The team used up to 94 qubits across two quantum processors for sampling, executing 9,200 circuits over more than 100 hours and collecting 1.3 billion measurement outcomes. The quantum data was then processed using powerful classical systems, including Japan's Fugaku supercomputer.
"This result is one of those things you dream about," said Dr. Kenneth Merz, who led the study.
The approach builds on a method that breaks large molecules into smaller, manageable clusters. Classical computers solve simpler regions, while quantum systems tackle the most entangled and computationally difficult parts. The results are then recombined to produce an overall picture of the molecule.
Researchers also introduced improvements to both classical and quantum techniques. One key step involved refining how the system identifies which parts of a molecule need detailed quantum treatment, reducing the overall computational cost.
Another advance came from a new quantum algorithm that improves how relevant electronic configurations are identified. This helps the system focus on the most important parts of a molecule's behavior while ignoring less useful data.
Despite the progress, the method does not yet outperform the best classical approaches. However, it demonstrates that quantum systems can already contribute to meaningful scientific problems, particularly when integrated with existing computing infrastructure.
"If we want another order-of-magnitude-or-two bump, quantum computing is probably the way to go," Merz said.
The findings suggest that hybrid quantum-classical workflows could become a practical tool for chemistry, especially as quantum hardware continues to improve. Future systems are expected to handle even larger and more complex molecules with greater accuracy.
The potential applications are significant. More accurate simulations could speed up drug discovery, improve materials design, and reduce the need for costly laboratory experiments.
The research highlights how combining quantum processors with classical computing resources may define the next phase of high-performance computing, offering a path toward solving problems that are currently out of reach.
