Speaker
Description
Molecular docking is a computational technique for predicting the binding of a small molecule to a protein. Here, we tackle molecular docking through quantum computing. To do so, we must define metrics for its resolution. We present a formulation to identify the best ligand orientation and conformation within a protein binding site as a multi-objective optimisation problem. Therefore, we want to maximise feasible interactions while minimising the penalties due to geometrical conflicts. We formulate molecular docking as a maximum weighted independent set (MWIS), a quadratic unconstrained binary optimisation (QUBO) problem [1]. QUBO problems can be found in drug discovery, logistics, or the design of new materials. These combinatorial problems, where the solution space grows exponentially, are the best candidates for quantum architectures. We solve the MWIS problem with DWAVE’s digital annealing (DA). Figure 1 displays the interaction capabilities of eight ligands. Our formulation can determine both the interaction capabilities and the interaction locations of different ligands and an enzyme related to cardiovascular diseases.
[1] Garrigues, Mathieu, et al. "A Scalable Heuristic for Molecular Docking on Neutral-Atom Quantum Processors." arXiv preprint arXiv:2508.18147 (2025).
Acknowledgements: We thank CESGA and, in particular, the Quantum Computing group members for their technical assistance, feedback, and the stimulating intellectual environment they provide. This work was funded through the project COCUMO (COmputación CUántica para MOléculas de interés) by the Ministerio de Industria y Comercio in collaboration with Hijos de Rivera S.A. Simulations are performed using the Finisterrae III Supercomputer, funded by the project CESGA-01 FINISTERRAE III.