Files
Abstract
The fields of lipidomics and metabolomics have been greatly impacted by the growing popularity surrounding ion mobility-mass spectrometry. Due to the sample complexity found in these studies, implementing a tool that can distinguish structural differences in thousands of compounds is invaluable to increasing the efficiency and accuracy of feature identification. However, generating large datasets makes data processing and analysis more challenging and lends to the need for generating workflows that can handle the intricacies involved. For this reason, there will always be room for new data collection techniques, analysis software, and spectral libraries upon which to be improved. The work described herein addresses some of these workflow improvements in regard to traveling wave-ion mobility specifically.First, we describe lipid isomer differentiation by identifying fatty acid double bond positions in standards and a total polar tissue extract using ion mobility shift reagents. To do this we use the Paternò-Büchi reaction and halogenated acetophenone compounds to derivatize unsaturated lipids, followed by collision induced dissociation to yield diagnostic ions indicative of the originating double bond position. Secondly, we developed an open-
source software, dubbed MOCCal, that can calculate calibrated traveling wave CCS values for lipids, small molecules, and singly-, doubly-, and triply-charged peptides. To showcase MOCCal’s metrics of accuracy and reproducibility, LC-IM-MS and FI-IM-MS experiments were performed with the well-characterized NIST SRM 1950. Lastly, we constructed a high-throughput, multi-omic workflow for the lipid and metabolite profiling of a variety of antibiotic resistant and susceptible bacterial strains. We were able to successfully replicate bacterial profiling of a variety of ESKAPE pathogen isolates conducted with 17-minute HILIC runs using 2-minute flow injection runs without a column and link the resulting multi-omic profiles to antibiotic resistance phenotypes. Our results are promising for developing a multi-omic based classification model for unknown bacterial strain identification and antibiotic susceptibility predictions.