Automated Cyclic Ion Mobility–Mass Spectrometry MS/MS Workflow and Data Analysis for Shotgun Lipidomics

Significance of the topic

Lipidomics is a key analytical approach for understanding cellular phenotypes, signaling, and metabolic states. Improved mass spectrometry-based lipid profiling enables characterization of cellular heterogeneity, cell–cell interactions, and dynamic lipid remodeling down to small sample amounts and single cells. However, shotgun lipidomics by direct infusion suffers from spectral complexity caused by isobaric and isomeric lipid species that cannot be resolved by m/z alone. Integrating ion mobility separations, specifically cyclic ion mobility (cIM), adds a drift-time dimension that separates lipids by shape/size and facilitates cleaner precursor and fragment spectra, increasing confidence in lipid identification and quantitation.

Objectives and study overview

This work demonstrates the integration of cyclic ion mobility-mass spectrometry (cIM-MS) into an automated direct-infusion (shotgun) lipidomics workflow. The main objectives were to (1) implement an acquisition strategy combining cIM-MS (MS1) and quadrupole-selected cIM MS/MS (Q-cIM-MS/MS), (2) develop automated data-processing capable of handling two-dimensional data (m/z and drift time) for both MS1 and MS2, and (3) demonstrate proof-of-principle improvements in resolving isobars/isomers and in quantitative coverage of lipid classes.

Methodology and data processing

Sample preparation and acquisition:

• Biological extracts were prepared by a customized MTBE (methyl‑tert‑butyl ether) extraction protocol with five independent extractions per sample set; extracts were diluted in MS solvent prior to analysis.
• Direct infusion (no LC) experiments were performed using nano‑ESI via a Triversa NanoMate under positive-ion mode.
• Acquisition modes included HDMS (cIM-MS) for data-rich MS1 maps and HDMS/MS (Q-cIM-MS/MS) for drift‑time–resolved MS/MS in v‑mode.

Data processing pipeline and algorithms:

• Multidimensional peak detection was implemented to find features in the two-dimensional space (m/z vs drift time) for both precursor and fragment ions.
• Drift times were converted to collision cross section (CCS)-relevant metrics using Python-based tools (pandas, numpy) integrated with ApexRT for peak detection.
• Lipid identification employed prototype adaptations of LipidXplorer and lxPostman to support multidimensional MasterScan generation and MFQL-based annotation.
• Automated generation of 2D feature lists enabled downstream identification, QC, and quantitation workflows.

Instrumentation used

Key instrumentation and software used in the study included:

• SELECT SERIES Cyclic IMS platform (Waters Corporation) for cyclic ion mobility separations.
• Nano‑electrospray using Triversa NanoMate (Advion); typical settings reported: ~1.1 kV at 1.0 psi and source temperature ~150 °C.
• Acquisition and review: MassLynx v4.2 and DriftScope 3.0.
• Peak detection and processing: ApexRT 1.17; Python libraries (pandas, numpy) for drift-time/CCS computations.
• Lipid identification: Prototype LipidXplorer 1.5 and lxPostman extensions to support multidimensional data.

Main results and discussion

Resolution and spectral clarity:

• cIM added an orthogonal separation axis that resolved isobaric and isomeric lipid species that overlapped in m/z space, demonstrated by drift-time–resolved arrival time distributions (ATDs) for selected precursors (e.g., m/z 759.5 and m/z 766.5).
• Q-cIM-MS/MS produced MS/MS spectra with reduced fragment overlap, enabling clearer assignment of diagnostic fragments (for example, separation of PC and ether‑linked PC species in 2D maps and IM‑resolved fragment ATDs for SM species).

Coverage, throughput, and quantitation:

• In a proof-of-principle positive‑ion shotgun experiment, 166 lipid species from eight lipid classes were quantified.
• High-throughput acquisition was demonstrated: an experimental design collected ~250 Q-cIM-MS/MS events within 6 minutes using a direct-infusion setup.
• Two-dimensional (m/z, IM) representation improved visualization and quantitation of classes such as triacylglycerols (TG) and phosphatidylcholines (PC), reducing background from isobaric interferences.

Data handling and automation:

• Multidimensional peak picking and automated feature-list generation enabled semi-supervised processing and potential scaling to larger cohorts.
• Prototype software adaptations were necessary to perform MasterScan generation and MFQL-based annotation in a multidimensional context; these underline current informatics gaps for routine deployment.

Benefits and practical applications

The integrated Q-cIM-MS/MS shotgun workflow provides several practical advantages:

• Improved isomer/isobar separation without chromatography, enabling higher throughput sample processing suitable for large cohorts or rapid profiling.
• Cleaner MS/MS spectra for more confident structural annotation of lipids, particularly for classes with overlapping diagnostic fragments (e.g., PC, PC‑O, SM).
• Applicability to limited sample inputs or high-throughput screening where LC times are prohibitive, and potential extension toward single-cell lipidomics with further sensitivity gains.

Limitations and considerations

Key limitations observed or implicit in this proof-of-principle study:

• Experiments were single‑pass cIM runs; higher resolving power could require multi‑pass cIM or optimized modulation at the cost of acquisition complexity.
• Quantitative accuracy and dynamic range in direct‑infusion formats depend on ion suppression and the effectiveness of IMS separation to reduce interferences; absolute quantitation requires careful internal-standard strategies and calibration.
• Current informatics relies on prototype software and semi‑supervised workflows; robust, validated pipelines and standardized CCS libraries are needed for routine, large-cohort deployment.

Future trends and potential applications

Expected developments and opportunities for applying Q-cIM-MS/MS in lipidomics include:

• Broader adoption of automated 2D feature lists, integrated CCS libraries, and community standards to improve identification confidence across labs.
• Extensions to both positive and negative ion modes with inclusion-list strategies to increase class coverage and targeted MS/MS throughput.
• Application to larger clinical or biological cohorts, combining speed of shotgun workflows with cIM-driven specificity to profile disease-related lipid signatures.
• Integration of machine‑learning methods for multidimensional peak detection and spectral annotation to accelerate interpretation and reduce manual review.
• Higher-resolution IM (multi‑pass cIM) and improved fragmentation schemes to resolve challenging isomeric lipids and provide deeper structural information (double-bond position, sn‑position) when complemented by targeted chemistries.

Conclusions

This study demonstrates that cyclic ion mobility coupled to quadrupole‑selected MS/MS (Q-cIM-MS/MS) materially improves the analytical performance of direct‑infusion shotgun lipidomics by resolving isobaric and isomeric interferences and producing cleaner MS/MS spectra. The developed acquisition schemes and prototype multidimensional data‑processing pipeline provide a scalable framework for higher‑throughput lipid profiling with improved identification confidence. To transition from proof‑of‑principle to routine use will require further advances in informatics, standardized CCS references, and validation across diverse sample types and larger cohorts.

Reference

1. Giles, K. et al. A Cyclic Ion Mobility–Mass Spectrometry System. Analytical Chemistry. 2019.
2. Matyash, V. et al. Lipid extraction by methyl‑tert‑butyl ether for high‑throughput lipidomics. Journal of Lipid Research. 2008.
3. Herzog, R. et al. A novel informatics concept for high‑throughput shotgun lipidomics based on the molecular fragmentation query language. Genome Biology. 2011.
4. lxPostman: lifs-tools.org/tools/lxpostman.html