Efficient Quantum Protein Structure Prediction with Problem-Agnostic Ansatzes
Hanna Linn, Ruihao Li, Holden, Alexander, Abdullah Ash Saki, Frank P. DiFilippo, Tomas Radivoyevitch, Daniel J Blankenberg, Laura García-Álvarez, Göran Johansson
Accurately predicting protein structures from amino acid sequences remains a fundamental challenge in computational biology, with profound implications for understanding biological functions and enabling structure-based drug discovery. Quantum computing approaches based on coarse-grained lattice models combined with variational algorithms have been proposed as an initial step towards predicting protein structures using quantum computers. In this work, we introduce a more efficient quantum protein structure prediction workflow that bypasses the need for explicit Hamiltonian construction by employing a problem-agnostic ansatz. The ansatz is trained to minimize an energy-based cost function that can be efficiently computed on classical computers, eliminating the need for ancillary qubits and reducing circuit depth compared to previous Hamiltonian-based methods. This enables a more scalable approach for larger proteins and facilitates the inclusion of higher-order interactions, previously hard to achieve in quantum approaches. We validate our method by benchmarking a hardware-efficient ansatz on a large set of proteins with up to 26 amino acids, modeled on the tetrahedral, body-centered cubic, and face-centered cubic lattices, incorporating up to second-nearest-neighbor interactions. We assess the performance on both a noise-free simulator and the ibm_kingston quantum computer using a set of distinct metrics to probe different aspects of the prediction quality. These experiments push the boundaries of quantum methods for protein structure prediction, targeting sequences that are longer than those typically addressed in prior studies. Overall, the results highlight the scalability and versatility of our approach, while also identifying key areas for improvement to inform future algorithm development and hardware advancements.
Accurately predicting protein structures from amino acid sequences remains a fundamental challenge in computational biology, with profound implications for understanding biological functions and enabling structure-based drug discovery. Quantum computing approaches based on coarse-grained lattice models combined with variational algorithms have been proposed as an initial step towards predicting protein structures using quantum computers. In this work, we introduce a more efficient quantum protein structure prediction workflow that bypasses the need for explicit Hamiltonian construction by employing a problem-agnostic ansatz. The ansatz is trained to minimize an energy-based cost function that can be efficiently computed on classical computers, eliminating the need for ancillary qubits and reducing circuit depth compared to previous Hamiltonian-based methods. This enables a more scalable approach for larger proteins and facilitates the inclusion of higher-order interactions, previously hard to achieve in quantum approaches. We validate our method by benchmarking a hardware-efficient ansatz on a large set of proteins with up to 26 amino acids, modeled on the tetrahedral, body-centered cubic, and face-centered cubic lattices, incorporating up to second-nearest-neighbor interactions. We assess the performance on both a noise-free simulator and the ibm_kingston quantum computer using a set of distinct metrics to probe different aspects of the prediction quality. These experiments push the boundaries of quantum methods for protein structure prediction, targeting sequences that are longer than those typically addressed in prior studies. Overall, the results highlight the scalability and versatility of our approach, while also identifying key areas for improvement to inform future algorithm development and hardware advancements.
Abstract Predicting the three-dimensional structure of a protein from its primary sequence of amino acids is known as the protein folding problem. Due to the central role of proteins’ structures in chemistry, biology and medicine applications, this subject has been intensively studied for over half a century. Although classical algorithms provide practical solutions for the sampling of the conformation space of small proteins, they cannot tackle the intrinsic NP-hard complexity of the problem, even when reduced to the simplest Hydrophobic-Polar model. On the other hand, while fault-tolerant quantum computers are beyond reach for state-of-the-art quantum technologies, there is evidence that quantum algorithms can be successfully used in noisy state-of-the-art quantum computers to accelerate energy optimization in frustrated systems. In this work, we present a model Hamiltonian with O(N4) O ( N 4 ) scaling and a corresponding quantum variational algorithm for the folding of a polymer chain with N monomers on a lattice. The model reflects many physico-chemical properties of the protein, reducing the gap between coarse-grained representations and mere lattice models. In addition, we use a robust and versatile optimization scheme, bringing together variational quantum algorithms specifically adapted to classical cost functions and evolutionary strategies to simulate the folding of the 10 amino acid Angiotensin on 22 qubits. The same method is also successfully applied to the study of the folding of a 7 amino acid neuropeptide using 9 qubits on an IBM 20-qubit quantum computer. Bringing together recent advances in building gate-based quantum computers with noise-tolerant hybrid quantum-classical algorithms, this work paves the way towards accessible and relevant scientific experiments on real quantum processors.
In the convergence of quantum computing and life science, we explore protein structure prediction and design on near-term intermediate-scale quantum devices. We investigate the algorithmic and resource constraints of today’s quantum computers, aiming to assess their potential in solving biologically relevant problems. We describe key variational quantum algorithms, including the problem-informed Quantum Approximate Optimization Algorithm and the problem-agnostic Hardware-Efficient Ansatz. Additionally, quantum walks are examined. The computationally complex coarse-grained lattice models in protein structure prediction and design are discussed. Quantum algorithms are then applied to these models to address the utility and limitations of today’s quantum computers. The thesis critically evaluates the limitations of quantum methods in comparison to classical approaches, highlighting the trade-offs between resource requirements in today’s quantum devices and the performance of quantum algorithms. Through this interdisciplinary investigation, the work contributes to understanding how quantum algorithms may advance computational biology in today’s quantum computing landscape.
Protein structure prediction (PSP) predicts the native conformation for a given protein sequence. Classically, the problem has been shown to belong to the NP-complete complexity class. Its applications range from physics, through bioinformatics to medicine and quantum biology. It is possible however to speed it up with quantum computational methods, as we show in this paper. Here we develop a fast quantum algorithm for PSP in three-dimensional hydrophobic-hydrophilic model on body-centered cubic lattice with quadratic speedup over its classical counterparts. Given a protein sequence of n amino acids, our algorithm reduces the temporal and spatial complexities to, respectively, O(2\frac n2) and O(n2logn). With respect to oracle-related quantum algorithms for the NP-complete problems, we identify our algorithm as optimal. To justify the feasibility of the proposed algorithm we successfully solve the problem on IBM quantum simulator involving 21 and 25 qubits. We confirm the experimentally obtained high probability of success in finding the desired conformation by calculating the theoretical probability estimations.
Protein folding, which traces the protein three-dimensional (3D) structure from its amino acid sequence, is a half-a-century-old problem in biology. The function of the protein correlates with its structure, emphasizing the need to study protein folding to understand the cellular and molecular processes better. While recent AI-based methods have shown significant success in protein structure prediction, their accuracy diminishes with proteins of low sequence similarity. Classical simulations face challenges in generating extensive conformational samplings. In this work, we develop a novel turn-based encoding algorithm with more significant degrees of freedom that successfully runs on a gate-based quantum computer and predicts the structure of proteins of varied lengths utilizing up to 114 qubits (IBM hardware). To make the problem tractable in quantum computers, the protein sequences were described with the simplistic HP model (H = hydrophobic residues, P = polar residues). The proposed formulation successfully captures the so-called nucleation step in protein folding, the hydrophobic collapse, that brings the hydrophobic residues to the core of the protein.
Accurate prediction of protein active-site structures remains a central challenge in structural biology, especially for short and flexible peptide fragments where conventional and simulation-based methods often fail. Here, we present a quantum computing framework designed for utility-level quantum processors to address this problem. Starting from an amino acid sequence, we cast structure prediction as a ground-state energy minimization task using the Variational Quantum Eigensolver (VQE). Amino acid connectivity is represented on a tetrahedral lattice, and steric, geometric, and chirality constraints are encoded into a problem-specific Hamiltonian expressed as sparse Pauli operators. A two-stage architecture separates energy estimation from measurement decoding, enabling noise mitigation under realistic device conditions. We evaluate the method on 23 real protein fragments from the PDBbind dataset and 7 fragments from therapeutically relevant proteins, executing all experiments on the IBM-Cleveland Clinic quantum processor. Structural predictions are benchmarked against AlphaFold3 (AF3) and classical simulation-based approaches using identical postprocessing and docking procedures. Our quantum framework outperforms both AF3 and classical baselines in Root-Mean-Square Deviation (RMSD) and docking efficacy, demonstrating a practical end-to-end pipeline for biologically relevant structure prediction on real quantum hardware and highlighting its engineering feasibility for near-term quantum devices.
Quantum heuristics have shown promise in solving various optimization problems, including lattice protein folding. Equally relevant is the inverse problem, protein design, where one seeks sequences that fold to a given target structure. The latter problem is often split into two steps: (i) searching for sequences that minimize the energy in the target structure, and (ii) testing whether the generated sequences fold to the desired structure. Here, we investigate the utility of variational quantum algorithms for the first of these two steps on today’s noisy intermediate-scale quantum devices. We focus on the sequence optimization task, which is less resource-demanding than folding computations. We test the quantum approximate optimization algorithm and variants of it, with problem-informed quantum circuits, as well as the hardware-efficient ansatz, with problem-agnostic quantum circuits. While the former approach, with a careful choice of mixer, yields good results in noiseless simulations (success probability ≥ 0.95 in all instances), its performance drops drastically under noise. With the problem-agnostic circuits, which are more compatible with hardware constraints, improved performance is observed in noisy simulations, compared to their problem-informed counterparts. When running the problem-agnostic circuits on a real quantum device, with parameters taken from simulations, we obtain a significant success probability ( ≥ 0.2 ) in most instances.