A recent study by researchers from Cleveland Clinic and IBM has been published in the Journal of Chemical Theory and Computation. This research could be a significant step towards using quantum computing techniques for protein structure prediction. It marks the first peer-reviewed quantum computing paper from the Cleveland Clinic-IBM Discovery Accelerator partnership.
Scientists have long used computational methods to forecast protein structures,
Protein structures are important for determining how a protein functions and interacts with other molecules in the body. These structures play a crucial role in human health and disease.
Accurately predicting the structure of a protein can help researchers gain a better understanding of how diseases spread and develop more effective therapies. Bryan Raubenolt, Ph.D., a postdoctoral fellow at Cleveland Clinic, and Hakan Doga, Ph.D., a researcher at IBM, led a team to explore how quantum computing can enhance current methods.
In recent years, machine learning techniques have made significant advancements in predicting protein structure.Fiction. These approaches depend on using training data, which is a database of experimentally determined protein structures, to make predictions. This means that their predictions are limited by the number of proteins they have been trained to identify. This limitation can result in lower levels of accuracy when the programs or algorithms encounter a protein that is mutated or very different from the ones they were trained on, which is common with genetic disorders.
Another approach is to use simulations to mimic the physics of protein folding. Simulations enable researchers to explore the various potential shapes of a given protein and identify the most stable one. The most stable shape is crucial.The use of quantum computing is essential for drug design. The problem is that running these simulations on a classical computer becomes nearly impossible once the target protein reaches a certain size. It’s like trying to solve a larger Rubik’s cube – for a small protein with 100 amino acids, a classical computer would need an impossibly long time to explore all the potential outcomes, according to Dr. Raubenolt. To address these limitations, the research team combined quantum and classical computing methods. This approach could enable quantum algorithms to tackle the challenging areas.The search for advanced classical computing technology encompasses a variety of factors such as protein size, intrinsic disorder, mutations, and the physics involved in protein folding. The effectiveness of the framework was confirmed through the precise prediction of the folding of a small section of a Zika virus protein on a quantum computer, in comparison to state-of-the-art classical methods.
The initial outcomes of the quantum-classical hybrid framework surpassed those of both a classical physics-based approach and AlphaFold2. Despite the fact that AlphaFold2 is specifically designed for larger proteins, this nonetheless proves the framework’s ability to generate accurate models without extreme limitations.
Researchers relied heavily on extensive training data.
They utilized a quantum algorithm to initially simulate the fragment’s backbone’s lowest energy conformation, which is usually the most challenging step in the calculation process. Classical methods were then employed to interpret the quantum computer’s results, rebuild the protein with its sidechains, and fine-tune the structure using classical molecular mechanics force fields. This project demonstrates how problems can be broken down into components, with quantum computing techniques handling some parts and classical computing handling others.
Enhancing precision has been a key focus of the project.
Dr. Raubenolt highlights the project’s unique aspect, stating, “One of the most distinctive elements of this project is the involvement of various disciplines. Our team’s expertise encompasses computational biology, chemistry, structural biology, software and automation engineering, experimental atomic and nuclear physics, mathematics, as well as quantum computing and algorithm design. The collaboration of knowledge from each of these fields was vital in creating a computational framework capable of simulating a crucial process for human life.”
The team’s utilization of both classical and quantum computing methods marks a significant advancement in our comprehension of the subject.The focus of the research is on the prediction of protein structures and how they can impact the treatment and prevention of diseases. The team aims to continue developing and improving quantum algorithms that can forecast the structure of more complex proteins in the future.
Dr. Doga states, “This research represents a significant advancement in exploring the potential capabilities of quantum computing in predicting protein structures. Our objective is to create quantum algorithms that can accurately predict protein structures.”
Â
