Towards the simulation of large scale protein–ligand interactions on NISQ-era quantum computers
Fionn D. Malone, Robert M. Parrish, Alicia R. Welden, Thomas Fox, Matthias Degroote, Elica S. Kyoseva, Nikolaj Moll, Raffaele Santagati, Michael Streif
accuracy in the SAPT interaction energies. In SAPT(VQE), the quantum requirements, such as qubit count and circuit depth, are lowered by performing computations on the separate molecular systems. Furthermore, active spaces allow for large systems containing thousands of orbitals to be reduced to a small enough orbital set to perform the quantum portions of the computations. We benchmark SAPT(VQE) (with the VQE component simulated by ideal statevector simulators) against a handful of small multi-reference dimer systems and the iron center containing human cancer-relevant protein lysine-specific demethylase 5 (KDM5A).
accuracy in the SAPT interaction energies. In SAPT(VQE), the quantum requirements, such as qubit count and circuit depth, are lowered by performing computations on the separate molecular systems. Furthermore, active spaces allow for large systems containing thousands of orbitals to be reduced to a small enough orbital set to perform the quantum portions of the computations. We benchmark SAPT(VQE) (with the VQE component simulated by ideal statevector simulators) against a handful of small multi-reference dimer systems and the iron center containing human cancer-relevant protein lysine-specific demethylase 5 (KDM5A).
The computation of interaction energies on noisy intermediate-scale quantum (NISQ) computers appears to be challenging with straightforward application of existing quantum algorithms. For example, use of the standard supermolecular method with the variational quantum eigensolver (VQE) would require extremely precise resolution of the total energies of the fragments to provide for accurate subtraction to the interaction energy. Here we present a symmetry-adapted perturbation theory (SAPT) method that may provide interaction energies with high quantum resource efficiency. Of particular note, we present a quantum extended random-phase approximation (ERPA) treatment of the SAPT second-order induction and dispersion terms, including exchange counterparts. Together with previous work on first-order terms, this provides a recipe for complete SAPT(VQE) interaction energies up to second order. The SAPT interaction energy terms are computed as first-level observables with no subtraction of monomer energies invoked, and the only quantum observations needed are the the VQE one- and two-particle density matrices. We find empirically that SAPT(VQE) can provide accurate interaction energies even with coarsely optimized, low circuit depth wavefunctions from the quantum computer, simulated through ideal statevectors. The errors on the total interaction energy are orders of magnitude lower than the corresponding VQE total energy errors of the monomer wavefunctions.
We analyzed the performance of quantum algorithms for studying protein-ligand systems, using both simulator and real hardware. To achieve this, we selected thrombin and 5 ligands as a test case and employed a decomposition strategy using density matrix embedded theory, dividing the ligand systems into three fragments each. First, the energy of one of the fragments was calculated using the Variational Quantum Eigensolver (VQE) using a state vector simulator, while the energy of the remaining fragments was calculated using Couple Cluster Singles and Doubles (CCSD), with the protein treated as point charges located at the respective atomic sites. Through these noiseless simulations, we evaluated the results under ideal conditions using the state-of-the-art Unitary CCSD ansatz to validate the efficacy of the strategy in a controlled setting without quantum noise. Subsequently, we evaluated the impact of approximations in the ansatz, optimizer, and active space size, which are necessary to decrease the computational cost, in order to target real hardware during the current Noisy Intermediate-Scale Quantum era. Finally, we performed the VQE calculations using a superconducting quantum computer developed by the RIKEN RQC-Fujitsu Collaboration Center and also analyzed the noise effect through simulations. The results demonstrated that the use of quantum algorithms can enhance the binding energy correlation value, offering potential applications in the workflow of computer-aided drug design.
In recent years, the Variational Quantum Eigensolver (VQE) has emerged as one of the most popular algorithms for solving the electronic structure problem on near-term quantum computers. The utility of VQE is often hindered by the limitations of current quantum hardware, including, but not limited to, short qubit coherence times and low gate fidelities. These limitations become particularly pronounced when VQE is used along with deep quantum circuits, such as those required by the ”Unitary Coupled Cluster Singles and Doubles” (UCCSD) ansatz, often resulting in significant errors. To address these issues, we propose a low-depth ansatz based on parallelized Givens rotations, which can recover substantial correlation energy while drastically reducing circuit depth and two-qubit gate counts for an arbitrary active space (AS). Also, considering the current hardware architectures with low qubit counts, we introduce a systematic way to select molecular orbitals to define active spaces (ASs) that retain significant electron correlation. We validate our approach by computing bond dissociation profiles of water and strongly correlated systems, such as molecular nitrogen and oxygen, across various ASs. Noiseless simulations using the new ansatz yield ground-state energies comparable to those from the UCCSD ansatz while reducing circuit depth by 50–70%. Moreover, in noisy simulations, our approach achieves energy error rates an order of magnitude lower than that of UCCSD. Considering the efficiency and practical usage of our ansatz, we hope that it becomes a potential choice for performing quantum chemistry calculations on near-term quantum devices.
We present a large-scale experimental study of quantum-computing-based molecular simulation carried out on IQM's Sirius 24-qubit superconducting processor, utilizing up to 16 operational qubits. The work employs Sample-based Quantum Diagonalization (SQD) together with the Local Unitary Cluster Jastrow (LUCJ) ansatz to estimate ground-state energies for a set of benchmark molecules, including H2, LiH, BeH2, H2O, and NH3. In addition, we introduce a Linear-CNOT variant of the Unitary Coupled-Cluster Singles and Doubles (LCNot-UCCSD) ansatz within the SQD workflow, trading higher circuit depth for reduced classical preprocessing. A comparison between these ansätze is provided, clarifying their respective strengths, limitations, and suitability for near-term quantum hardware. We further explore potential energy landscapes through 1D scans for H2 and HeH+ using both STO-3G and 6-31G basis sets, and for LiH and BeH2 in STO-3G. Extending beyond this, we demonstrate the experimental construction of a full 2D potential energy surface for the water molecule on quantum hardware, mapped over a 32 × 32 grid in bond length and bond angle. To move beyond small benchmark systems, we combine SQD(LUCJ) with Density Matrix Embedding Theory (DMET) to compute active-space energies for a set of ligand-like molecules, as well as the pharmacologically relevant amantadine system. Across all studies, the majority of quantum-computed energies agree with reference FCI results, as well as with DMET-CASCI energies for embedded systems, to within chemical accuracy for the chosen basis sets. These results demonstrate the reliability of sample-based diagonalization approaches and underscore the potential of hybrid embedding strategies for extending quantum simulations to increasingly complex molecular systems, while also highlighting their practicality on current IQM quantum hardware.
A central problem in biophysics and computational drug design is accurate modeling of biomolecules. The current molecular dynamics simulation methods can answer how a molecule inhibits a cancerous cell signaling pathway, or the role of protein misfolding in neurodegenerative diseases. However, the accuracy of current force fields (interaction potential) limits the reliability of computer simulations. Fundamentally a quantum chemistry problem, here we discuss optimizing force fields using scalable ab initio quantum chemistry calculations on quantum computers and estimate the quantum resources required for this task. For a list of dipeptides for local parameterizations, we estimate the required number of qubits to be 1576 to 3808 with cc-pVTZ(-f) orbital basis and 88 to 276 with active space reduction. Using a linear depth ansatz with active-space reduction, we estimate a quantum circuit with a circuit depth of few thousands can be used to simulate these dipeptides. The estimated number of 100s of qubits and a few thousand long circuit depth puts the pharmaceutical application of near-term quantum processors in a realistic perspective.
Calculation of binding energies for protein-ligand molecular systems requires accurate treatment of the electronic structure, a quantum chemistry problem that scales exponentially on classical hardware, while current quantum hardware remains too noisy for the required circuit depths. This report presents a hybrid quantum-classical workflow performed on the Fujitsu FX700 ideal state-vector simulator using QARP that addresses two structural inefficiencies in quantum-sampling-based diagonalization workflows. First, we integrate the Linear Scaling CNOT UCCSD (LCNot-UCCSD) ansatz into the QSCI framework, replacing the O(N6) CCSD parameter initialization of the competing LUCJ ansatz approach with O(N4) MP2-amplitude initialization. Second, we introduce QSCI-RBM, a variant that replaces the configuration recovery of the SQD framework with a Restricted Boltzmann Machine (RBM) acting as a compact generative subspace expansion model. Both are evaluated on eight different molecules in STO-3G across 14 controlled artificial error levels with 100 independent runs each, validated on potential energy surface scans of the N2 molecule in cc-pVDZ, and embedded within DMET to treat the FDA-approved antiviral Amantadine (C10H17N, 11 DMET fragments) and the active region of the SARS-CoV-2 main protease complexed with its covalent inhibitor Carmofur (PDB: 7BUY, C15H28N4O5S, 10 fragments). To our knowledge, this is the first deployment of LCNot-UCCSD within QSCI on a quantum computing simulator, and the first DMET-QSCI(LCNot-UCCSD)-RBM application to an industry-relevant protein-ligand system. By utilizing a fraction of the classical computing resources required by the current state-of-the-art work by Cleveland Clinic, RIKEN, and IBM Quantum, this approach enables more efficient and economical drug discovery simulations for the industry.