Lipidomics: Processing a Complex Lipid Dataset with LipidSearch

Lipidomics can provide vital information essential for understanding a wide range of disease states, particularly in cancer and diabetes. Because of the complexities of the lipidome, accessing this information using untargeted liquid chromatography and tandem mass spectrometry (LC-MS/MS) presents a challenge. For the most comprehensive analyses of data, past researchers have been dependent upon complex software and large databases.

Recently, a team of scientists demonstrated the capacity of LipidSearch software (Thermo Scientific) as a strategy to overcome the obstacles associated with manually processing MS data from lipidomics studies.¹ In a 60-min liquid chromatography and mass spectrometry (LC-MS) run, the scientists were able to identify and quantify approximately 1,000 isomeric lipid species from human plasma using an experimental C30 ultra-high-performance liquid chromatography (UHPLC) column.

The research team performed LC-MS using three aliquots of human plasma provided by the NIST. For these experiments, the team relied upon a Dionex UltiMate 3000 Rapid Separation LC [RSLC] system and a Q Exactive HF hybrid quadrupole-Orbitrap mass spectrometer (both Thermo Scientific) .

The LipidSearch software enhances lipid identification by searching databases of precursor accurate masses and their predicted fragment ions. When hits are discovered, the software identifies the lipids by ranking them based on mass tolerance and matching them to the theoretical fragment ions and fraction of total MS2 intensity. Additionally, LipidSearch identifies the lipid species in each LC-dd-MS2 experiment, assessing them at sum composition (MS) and isomer (MS2) levels. The team obtained each lipid identification using a single, high-quality Orbitrap MS2 scan over four orders of concentration dynamic range. They identified potential lipid species using the predicted MS2 fragments for molecular species observed in positive or negative ion mode. Table 1 summarily provides, for each lipid sub-class, the numbers of lipid species.

Table 1. A summary of the number of lipid species for each lipid sub-class

Lipid Class	Unfiltered Species	Filtered Species	Reported Species
Che	22	19	18
DG	81	46	45
TG	665	468	452
PC	508	221	220
LPC	106	58	57
PE/LPE	53	33	32
PI	34	25	24
Cer	33	22	21
CerG	20	13	13
SM	138	92	91
Total	1,660	997	973

These results demonstrate that the LipidSearch software was able to process high-quality Orbitrap LC-MS2 untargeted lipidomics data. The software made it possible to comprehensively identify and quantify nearly 1,000 lipid species in a single LC-MS run; as another plus, the high mass accuracy in both MS (120K) and MS2 (30K) obtained with the Q Exactive HF instrument produced confident CV values below 15%, demonstrating excellent reproducibility. The scientists propose that this particular combination of software and instrumentation is ideal for analyzing data from lipidomics experiments.

Reference

1. Kiyonami, R., et al. (2015, May) “Processing of a complex lipid dataset for the NIST inter-laboratory comparison exercise for lipidomics measurements in human serum and plasma,” poster presented at the LIPID MAPS Annual Meeting, La Jolla, CA, May 12–13.