We have developed a novel molecular fingerprinting method based on persistent homology that can encode the geometrical and electronic structure of molecules for chemical applications. We have demonstrated its applicability on two different test cases. The first is related to non-covalent interactions between functional groups of materials and small gas molecules for environmental applications. The second is related to biomimetic catalysis via heme and non-heme Fe(IV)-oxo sites for C-H activation. For both cases, quantum chemical calculations were performed on a small number of molecules (50-100) for the generation of meaningful data. We have used these data in order to train a statistical model that includes the new fingerprinting method and machine learning algorithms. The trained models have been used for high-throughput virtual screening by predicting the properties of larger molecular databases (more than 100,000 entries) where quantum chemical data are not available. <br/>

Persistent homology for chemical applications: Story of birth and death

2019 Fall National Meeting

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.