| Abstract |
Owing to the recent development of the PM6 and PM6-DH+ semi-empirical methodologies, which belong to the neglect of diatomic differential overlap (NDDO) family, it was decided to carry out a study to assess whether these inexpensive and fast methodologies could be used with confidence to help solve mass spectrometry problems. As such, a report on the feasibility of using semi-empirical calculations to identify probable protonation sites in amino acids is presented. The optimised geometries obtained by the semi-empirical calculations were compared to several structures reported in the literature (obtained through high-level theoretical calculations) and reasonable agreement was found. The proton affinities derived from semi-empirical calculations were also compared with experimental data and benchmarked as well with predicted values from the literature (also obtained through high-level theoretical calculations). Semi-empirical calculations accurately predicted the most probable protonation site for all amino acids considered; thus leading to results comparable to those obtained by high-level calculations at an extremely low computational cost. Regarding the proton affinity estimates, deviations from the available experimental values are greater for the semi-empirical proton affinities than for those observed for high-level calculations. A statistical analysis of the data, at a confidence level of 99\%, also showed that the semi-empirical proton affinities were different from experimental values and high-level proton affinities were equivalent to experimental values. Nevertheless, the overall correlation of the semi-empirical data with experimental values is, at least, satisfactory. We believe therefore that this paper shows that semi-empirical methodologies, which are fast and inexpensive, can indeed solve mass spectrometry problems, or at least, facilitate a quicker path to the solution. |
