Automatic Generation of Local Vibrational Mode Parameters: From Small to Large Molecules and QM/MM Systems

J Phys Chem A. 2022 Dec 15;126(49):9313-9331. doi: 10.1021/acs.jpca.2c07871. Epub 2022 Dec 6.

Abstract

LModeAGen, a new protocol for the automatic determination of a nonredundant, complete set of local vibrational modes is reported, which is based on chemical graph concepts. Whereas local mode properties can be calculated for a selection of parameters targeting specific local modes of interest, a complete set of nonredundant local mode parameters is requested for the adiabatic connection scheme (ACS), relating each local vibrational mode with a normal mode counterpart, and for the decomposition of normal modes (CNM) in terms of local mode contributions, a unique way to analyze vibrational spectra. So far, nonredundant parameter sets have been generated manually following chemical intuition or from a set of redundant parameters in a trial-and-error fashion, which has hampered the study of larger systems with hundreds of parameters. LModeAGen was successfully applied for a test set of 11 systems, ranging from small molecules to the large QM (>100 atoms) subsystem of carbomonoxy-neuroglobin protein, described with a hybrid QM/MM method. The ωB97X-D/aug-cc-pVDZ, M06L/def2-TZVP, and QM/MM ωB97X-D/6-31G(d,p)/AMBER model chemistries were adopted for the description of the molecules in the test set. Our new protocol is an important step forward for a routine ACS and CNM analysis of the vibrational spectra of complex and large systems with hundreds of atoms, providing new access to important encoded electronic structure information.

MeSH terms

  • Proteins / chemistry
  • Quantum Theory*
  • Vibration*

Substances

  • Proteins