Sampling rare conformational transitions with a quantum computer.

Area of Science:

• Computational Chemistry
• Biophysics
• Quantum Computing

Background:

• Molecular Dynamics (MD) simulations are crucial for understanding biomolecular dynamics but computationally expensive for rare events.
• Path sampling methods improve efficiency but struggle to explore diverse conformational spaces.
• Simulating thermal fluctuations in metastable states requires significant computational resources.

Purpose of the Study:

• To develop a new algorithm for Molecular Dynamics (MD) simulations that integrates machine learning and quantum computing.
• To overcome limitations in efficiently generating paths for rare events in biomolecular systems.
• To provide a proof-of-concept for a hybrid quantum-classical approach in molecular simulations.

Main Methods:

• Derivation of a low-resolution, coarse-grained model of system dynamics using functional integral methods and machine learning.
• Utilizing a quantum annealer to sample transition paths of the coarse-grained model.
• Proof-of-concept application simulating a benchmark conformational transition with all-atom resolution on a D-Wave quantum computer.

Main Results:

• Demonstrated a novel algorithm integrating machine learning and quantum computing for MD simulations.
• Successfully sampled transition paths of a coarse-grained model using quantum annealing.
• Generated uncorrelated trajectories at each iteration, addressing a key challenge in path sampling.

Conclusions:

• The integrated machine learning and quantum computing approach offers a promising new paradigm for high-performance scientific computing in molecular simulations.
• Quantum annealing can efficiently generate uncorrelated trajectories, improving path sampling.
• This work provides a platform for future hybrid quantum-classical simulations of complex biomolecular systems.