Reviewed by Lexie CornerMay 31 2024
Recent findings from IBM and Cleveland Clinic researchers may pave the way for applying quantum computing techniques to protein structure prediction. These findings are published in the Journal of Chemical Theory and Computation. This publication represents the Cleveland Clinic-IBM Discovery Accelerator collaboration's first peer-reviewed paper on quantum computing.
The IBM Quantum System One at Cleveland Clinic. Image Credit: Cleveland Clinic.
For many years, researchers have used computational methods to predict protein structures. A protein folds into a structure that controls its molecular interactions and mode of action. These structures determine numerous facets of human health and illness.
Researchers can create more effective treatments by better understanding how diseases spread through precise protein structure predictions. Bryan Raubenolt, Ph.D., a Postdoctoral Fellow at the Cleveland Clinic, and Hakan Doga, Ph.D., a researcher at IBM, led a team to discover how quantum computing can enhance existing techniques.
Machine learning techniques have significantly advanced the prediction of protein structure in recent years. To make predictions, these techniques rely on training data, a database of protein structures determined through experimentation. This indicates that the number of proteins they have been trained to identify is a limitation. When programs or algorithms come across a protein that is mutated or significantly different from the ones they were trained on, as is frequently the case with genetic disorders, this can result in decreased accuracy levels.
A different approach is to model the physics involved in protein folding. Through simulations, scientists can examine multiple protein configurations and determine the most stable form, which is essential for drug design.
The challenge is that these simulations are nearly impossible on a classical computer beyond a certain protein size. In a way, increasing the size of the target protein is comparable to increasing the dimensions of a Rubik's cube. For a small protein with 100 amino acids, a classical computer would need the time equal to the age of the universe to exhaustively search all the possible outcomes.
Dr. Bryan Raubenolt, Postdoctoral Fellow, Cleveland Clinic
The research team combined quantum and classical computing techniques to get around these restrictions. Within this framework, quantum algorithms can tackle problems that current state-of-the-art classical computing finds difficult, such as the physics of protein folding, intrinsic disorder, mutations, and protein size.
The accuracy with which the framework predicted, on a quantum computer, the folding of a small fragment of the Zika virus protein, compared to the most advanced classical methods, served as validation.
The initial results of the quantum-classical hybrid framework outperformed both AlphaFold2 and a method based on classical physics. The latter shows that this framework can produce accurate models without directly relying on large amounts of training data, even though it is optimized for larger proteins.
The most computationally intensive part of the calculation usually involves modeling the lowest energy conformation for the fragment's backbone, which the researchers accomplish using a quantum algorithm. After that, classical methods were employed to translate the quantum computer's output, rebuild the protein along with its sidechains, and refine the structure one last time using force fields from classical molecular mechanics.
The project illustrates how problems can be broken down into smaller components for better accuracy. Some components can be addressed by quantum computing techniques, while classical computing methods can handle others.
Working across disciplines was crucial to creating this framework.
One of the most unique things about this project is the number of disciplines involved. Our team’s expertise ranges from computational biology and chemistry, structural biology, software, and automation engineering to experimental atomic and nuclear physics, mathematics, and, of course, quantum computing and algorithm design. It took the knowledge from each of these areas to create a computational framework that can mimic one of the most important processes for human life.
Dr. Bryan Raubenolt, Postdoctoral Fellow, Cleveland Clinic
The team’s combination of classical and quantum computing methods is essential for advancing our understanding of protein structures and how they impact our ability to treat and prevent disease. The team plans to continue developing and optimizing quantum algorithms that can predict the structure of larger and more sophisticated proteins.
This work is an important step forward in exploring where quantum computing capabilities could show strengths in protein structure prediction. Our goal is to design quantum algorithms that can find how to predict protein structures as realistically as possible.
Dr. Hakan Doga, Researcher, IBM
Journal Reference:
Doga, H., et al. (2024) A Perspective on Protein Structure Prediction Using Quantum Computers. Chemical Theory and Computation. doi.org/10.1021/acs.jctc.4c00067