Computational approach to structural identification of phospholipids using raw mass spectra from nanoflow liquid chromatography-electrospray ionization-tandem mass spectrometry

Abstract

A qualitative analysis tool (LiPilot) for identifying phospholipids (PLs), including lysophospholipids (LPLs), from biological mixtures is introduced. The developed algorithm utilizes raw data obtained from nanoflow liquid chromatography-electrospray ionization-tandem mass spectrometry experiments of lipid mixture samples including retention time and m/z values of precursor and fragment ions from data-dependent, collision-induced dissociation. Library files based on typical fragmentation patterns of PLs generated with an LTQ-Velos ion trap mass spectrometer are used to identify PL or LPL species by comparing experimental fragment ions with typical fragment ions in the library file. Identification is aided by calculating a confidence score developed in our laboratory to maximize identification efficiency. Analysis includes the influence of total ion intensities of matched and unmatched fragment ions, the difference in m/z values between observed and theoretical fragment ions, and a weighting factor used to differentiate regioisomers through data filtration. The present study focused on targeted identification of particular PL classes. The identification software was evaluated using a mixture of 24 PL and LPL standards. The software was further tested with a human urinary PL mixture sample, with 93 PLs and 22 LPLs identified.