Effective protocol for database similarity searching of heteronuclear single quantum coherence spectra

The nuclear magnetic resonance (NMR) chemical shift exquisitely describes the chemical environment of the atoms in a molecule. Here we describe methods that utilize this information as an experimental probe to match 2D NMR heteronuclear single quantum coherence (HSQC) spectra of pure, unknown compounds to a database of known compounds. We implemented and compared two different approaches for similarity searching of HSQC spectra. According to our findings, our new discrete self-adaptive differential evolution method performs better than the previously published shifted grid, multiple resolution approach. The new method is provided in detail and comparisons have been performed for a set of HSQC spectra. The similarity comparison involves a peak-to-peak matching of different spectra, followed by a selection criterion and ranking to establish a level of match.