Year: 2026
Type: preprint
Abstract: Hybrid simulations utilizing both quantum mechanics and molecular mechanics (QM/MM) are widely used to model complex problems, such as the electronic excitation of chromophores embedded in a protein environment. Its success requires a careful choice of the portion of the system to include in the QM region to obtain accurate results. Chemical intuition and/or proximity arguments are often used to select the QM region for excited state calculations aiming to compute absorption spectra, but such choices do not guarantee accuracy. We compare four schemes for selecting the minimal set of protein residues that must be treated quantum mechanically to reproduce a reference spectrum: distance analysis, charge shift analysis, point charge variation analysis, and subtractive analysis. Charge shift analysis is the most robust: in both systems studied it identifies the smallest QM region that reproduces the reference spectral shape and absorption maximum, at moderate computational cost, using a threshold that transfers between systems. Because the resulting regions are both accurate and minimal, converged excited state calculations become affordable at scale, laying a foundation for large, internally consistent datasets of excited state properties in extended systems. We release an open-source implementation that automates the full workflow, including residue selection and determination of the QM/MM boundary, so a converged QM region can be built for a new system without manual intervention.
Source: ChemRxiv
Language: English
Subfield: Atomic and Molecular Physics, and Optics
Field: Physics and Astronomy
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7222592997