Modeling Energy Flow in Molecules with New Computational Tools and Data Analysis

Dr. David Leitner of the University of Nevada, Reno is supported to develop computational methods and introduce machine learning (ML) to analyze large computational and experimental data sets to predict the nature and rate of energy flow in molecules. Catalysis of chemical reactions involving molecules attached to metal nanoparticles will be optimized from the information obtained about rates of energy flow in the reactants using the new computational tools and ML analysis. Dr. Leitner and his research group will also apply ML tools to analyze large data sets of rates of energy transfer computed for proteins, the molecular machines of the cell, to predict rates of energy flow during chemical reactions and the impact of mutations on protein function. The research that will be carried out involves developing ML tools for molecular science, materials chemistry and biophysical chemistry, all areas of great and expanding opportunity for the graduate and undergraduate students who will participate, with applications to nanotechnology, drug discovery and medicine. Dr. David Leitner and his research group will develop and apply new computational methods to study the impact of molecular composition and structure on vibrational energy dynamics. Large data sets of vibrational frequencies and anharmonic interactions for aromatic compounds will be computed and analyzed using machine learning (ML) tools, enabling prediction of the impact of chemical substitution on tuning and detuning resonances that facilitate energy flow. Catalyzing reactions by selectively exciting molecular vibrations has recently been achieved for nitrobenzenethiol decomposition via exciting plasmons in attached metal nanoparticles. Predicting the role of chemical substitution on energy transfer will inform how attaching different chemical groups will impact catalysis. Dr. Leitner and his research group will also apply ML tools to analyze large data sets of rates of energy transfer computed for protein receptors, with focus on G protein-coupled receptors (GPCRs), to identify structural and dynamic features of proteins that mediate energy flow and to predict the impact of mutations on GPCR function. Computer programs developed for this project and the data generated by the programs will be made available for wider use. Dr. Leitner will continue efforts to promote interactions between researchers in fields related to those of the proposed work by organizing meetings, and to develop interactions between researchers in different disciplines at the University of Nevada, Reno, serving on the Executive Committee of the university’s Center for Advanced Studies. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.