Elucidating the Role of Hydrogen Bonding in the Optical Spectroscopy of the Solvated Green Fluorescent Protein Chromophore: Using Machine Learning to Establish the Importance of High-Level Electronic Structure

Hydrogen bonding interactions with chromophores in chemicalandbiological environments play a key role in determining their electronicabsorption and relaxation processes, which are manifested in theirlinear and multidimensional optical spectra. For chromophores in thecondensed phase, the large number of atoms needed to simulate theenvironment has traditionally prohibited the use of high-level excited-stateelectronic structure methods. By leveraging transfer learning, weshow how to construct machine-learned models to accurately predictthe high-level excitation energies of a chromophore in solution fromonly 400 high-level calculations. We show that when the electronicexcitations of the green fluorescent protein chromophore in waterare treated using EOM-CCSD embedded in a DFT description of the solventthe optical spectrum is correctly captured and that this improvementarises from correctly treating the coupling of the electronic transitionto electric fields, which leads to a larger response upon hydrogenbonding between the chromophore and water.