Poster Presentation 24th Annual Lorne Proteomics Symposium 2019

Improved lipid annotation utilizing positive and negative ion MS2 / MS3 HCD and CID spectra  (#129)

Amanda Souza 1 , David A Peake 1 , Reiko Kiyonami 1 , Daniel Gachotte 2 , Gavin Reid 3 , Yasuto Yokoi 4 , Andreas Huhmer 1
  1. Thermo Fisher Scientific, San Jose, California, United States
  2. Corteva Agriscience, Indianapolis, Indiana, United States
  3. The University of Melbourne, Parkville, Victoria, Australia
  4. Mitsui Knowledge Industry, Atago, Tokyo, Japan

Introduction

We present improvements to Thermo Scientific™ LipidSearch software. New algorithms were introduced to reduce false positives, improve quantitation and automate searching of LC-MSn data obtained by higher collisional energy (HCD) and linear ion trap collisional induced dissociation (CID). The use of LC-MSn is applied for more complete characterization of triacylglycerols in total lipid extracts.

 

Methods

Bovine liver and insect larvae extracts were analyzed by data dependent LC/MS2 and MS3 using a Thermo Scientific™ Orbitrap ID-X™ Tribrid™ mass spectrometer. The MSn data were searched against the m/z of selected lipid precursor ions and their predicted product ions and neutral losses. Each lipid adduct ion is ranked by mass tolerance, match to the predicted fragmentation and fraction of total MS-MS intensity.

 

Results

The number of lipid species annotated in each experiment was assessed at the sum composition (MS) and isomer (MSn) levels. LC-dd-MS2/MS3 spectra for potential lipid species were annotated separately from positive and negative ion adducts and then merged into a single lipid result.  This approach provides lipid annotation that reflects the appropriate level of MS2/MS3 product ions and neutral losses from giving higher confidence in lipid annotations. Results were filtered by the minimum number of data points, signal-to-noise ratio, adduct ion, match score, ID quality, and coefficient of variation from replicate injections. Compared to the results generated only from dd-MS2 HCD, a combination of HCD and CID LC-MSn gave significantly higher quality lipid identifications.  From bovine liver, the number of PC and TG results with complete acyl chain information improved by 12% to 23%, respectively.

 

Conclusion

Lipid annotation is often over-reported because the current software/database approaches match mass spectral evidence to exact lipid structures [1]. The lipidomics community is working towards reporting the correct level of annotation based upon the available mass spectral information.  Utilizing a combination of HCD MS2 and intelligent product ion or neutral loss triggered CID MS2/MS3 data improves lipid annotation available within a single LC/MSn experiment.

  1. Rustam Y. H., Reid, G. E. Anal. Chem., 90(1), 374-397 (2018).