Influence of Lipid Fragmentation in the Data Analysis of Imaging Mass Spectrometry Experiments

Imaging mass spectrometry (IMS) is becoming an essential technique in lipidomics. Still, many questions remain open, precluding it from achieving its full potential. Among them, identification of species directly from the tissue is of paramount importance. However, it is not an easy task, due to the abundance and variety of lipid species, their numerous fragmentation pathways, and the formation of a significant number of adducts, both with the matrix and with the cations present in the tissue. Here, we explore the fragmentation pathways of 17 lipid classes, demonstrating that in-source fragmentation hampers identification of some lipid species. Then, we analyze what type of adducts each class is more prone to form. Finally, we use that information together with data from on-tissue MS/MS and MS³ to refine the peak assignment in a real experiment over sections of human nevi, to demonstrate that statistical analysis of the data is significantly more robust if unwanted peaks due to fragmentation, matrix, and other species that only introduce noise in the analysis are excluded.