A Guide for Optimization of Data‑Dependent Acquisition Settings in Metabolomics

Optimizing Data‑Dependent Acquisition (DDA) for Untargeted Metabolomics on the Agilent Revident LC/Q‑TOF — Technical Overview Summary

Significance of the topic
High‑confidence metabolite annotation in untargeted LC‑MS workflows is limited chiefly by incomplete and low‑quality MS/MS coverage. Optimizing data‑dependent acquisition (DDA) strategies increases the number of informative fragmentation spectra per injection and improves library matching confidence, which accelerates metabolite identification and downstream biological interpretation. The combination of robust HILIC separation, informed precursor selection, and iterative acquisition strategies yields richer MS/MS datasets for discovery metabolomics.

Objectives and overview of the study
The work presents a practical guide to optimize key DDA parameters on an Agilent Revident LC/Q‑TOF running MassHunter DA (v12.1+). Main aims were to: (1) compare Auto MS/MS and the newer Directed MS/MS modes; (2) define optimal settings for precursor thresholds, MS1/MS2 acquisition rates, number of precursors per cycle, active exclusion rules, and purity/centroiding; and (3) demonstrate exclusion‑list strategies and iterative injection logic to boost MS/MS coverage while balancing spectral quality for library matching.

Methodology and instrumentation
Sample preparation and chromatography

Biological matrix: human plasma extracted using Agilent Captiva EMR cartridges after protein precipitation (MeOH/EtOH 1:1), followed by reconstitution in ACN/MeOH/water (8:1:1) and spiking with a mixed standard (final ~6 µM).

Comparison sample: a rapid RP‑C18 plant extract method (ZORBAX Eclipse Plus C18, 2.1 × 50 mm) used for evaluating exclusion strategies.

Primary separation: robust HILIC workflow (Agilent InfinityLab Poroshell 120 HILIC‑Z, 2.1 × 150 mm, 2.7 µm), 23‑min gradient, column at 15 °C, autosampler 5 °C, mobile phase A = 20 mM ammonium acetate pH 9.3 + 5 µM medronic acid, B = ACN. Typical injection 3 µL; baseline peak width ~10 s.

Mass spectrometry and software

Instrument: Agilent Revident LC/Q‑TOF running MassHunter Data Acquisition for TOF/Q‑TOF LC/MS v12.1; data analyzed with MassHunter Qualitative Analysis v12.0/13.0.

Ionization and source settings (negative ESI): gas temp 225 °C, drying gas 9 L/min, nebulizer 30 psi, sheath gas temp 375 °C, sheath gas 12 L/min, capillary 3,000 V, nozzle 500 V, fragmentor 100 V, skimmer 45 V, instrument mode 1,700 fragile.

Fragmentation settings: three collision energies (10, 20, 40 V) used to match entries in an in‑house HILIC spectral library (558 metabolites with RT and MS/MS at these energies).

Key acquisition parameters and rationale

MS1 acquisition: typical starting point 6 Hz to ensure sufficient points across a ~10 s peak (≈9 points).

MS2 acquisition rates: evaluated 5, 25, and 50 Hz. Faster MS2 reduces cycle time (higher coverage) but lowers the number of summed transients and can remove low‑intensity fragments, reducing library match scores.

Precursors per cycle: recommended starting at 3 for HILIC with three collision energies; increasing precursors per cycle improves duty‑cycle efficiency when MS2 rate is high.

Active exclusion: exclude after 1–2 MS/MS events and release after ~one‑third to one‑half of the chromatographic peak width to sample apex and tails while maximizing diversity of sampled precursors.

Purity and isolation: narrow quadrupole isolation (~1.3 m/z), purity stringency set to 100% with a 30% cutoff to limit chimeric spectra.

Variable acquisition rate: can adapt MS2 transient count to precursor abundance to reach a target total ion count (recommended target ≥30,000) but note implementation limits (minimum transient floor ~200 in v12.1).

Data storage/centroiding: set profile storage threshold to zero (important for SureMass conversion). Example centroid thresholds: relative 1×10⁻⁴%, absolute 100 for MS1 and 5 for MS2 in this study.

Main results and discussion
Exclusion list strategies and MS/MS coverage

Four exclusion strategies were compared on a plant extract: (A) minimal reference mass exclusion, (B) static exclusion derived from a blank injection, (C) Directed MS/MS using a preferred list, and (D) dynamic iterative exclusion (iterative Auto MS/MS building an exclusion list across injections).

MS/MS coverage (single injection, plant extract): (A) reference only – 12%; (B) static blank exclusion – 29%; (C) Directed MS/MS – 27%; (D) dynamic iterative exclusion – 41%. The iterative dynamic exclusion was most effective and required least manual curation.

Trade‑offs between acquisition rate and spectral quality

Iterative Auto MS/MS at the highest MS2 acquisition rate (50 Hz) with three iterations delivered the maximum MS/MS coverage (~90%) across the plasma dataset.

However, higher MS2 rates reduced library matching quality: for Auto MS/MS average Library Scores were 69 (5 Hz), 54 (25 Hz), and 46 (50 Hz). Fraction of compounds with Library Score >85 fell from 38% (5 Hz) to 13% (50 Hz).

Directed MS/MS tended to yield higher single‑injection coverage than single‑injection Auto MS/MS at low‑to‑medium MS2 rates, making it valuable for semi‑targeted analyses focused on statistically significant features.

Practical recommendations resulting from the study

Use a robust blank to build an exclusion list as a minimal measure; apply iterative exclusion to maximize coverage with minimal manual effort.

Start with MS1 = 6 Hz, MS2 = 10–25 Hz and 3 precursors per cycle for HILIC; increase MS2 to 25–50 Hz and the number of precursors per cycle if peak widths and ion abundances permit.

Tune active exclusion (exclude after 1 spectrum; release after ~1/3–1/2 peak) and set precursor thresholds ~3× above noise to trigger early on peaks.

Consider variable acquisition rate and target ion counts (≥30,000) for high‑abundance precursors, but monitor impact on low‑abundance fragmentation and resultant library scores.

Preserve profile data (profile threshold = 0) if using SureMass or other algorithms that require raw profiles.

Benefits and practical applications
Optimized DDA methods described here enable substantially higher MS/MS coverage and better prioritization of biologically relevant precursors. This improves confidence in spectral library matches and expedites metabolite annotation workflows in discovery studies, biomarker research, and semi‑targeted follow‑up experiments. Directed MS/MS supports targeted interrogation of statistically significant features, while iterative Auto MS/MS maximizes total coverage for broad discovery work.

Future trends and applications

• Deeper automation of exclusion/preferred list generation from initial MS1 batches (integration with statistical pipelines) will streamline semi‑targeted studies and reduce manual curation.
• Advances in adaptive acquisition algorithms (real‑time prioritization, smarter variable acquisition) could further improve the balance between coverage and spectral quality.
• Improved spectral libraries with multi‑collision energy entries and community‑driven curation will increase identification confidence, particularly for low‑abundance metabolites.
• Integration with ion mobility, orthogonal separation, and enhanced informatics (AI‑assisted spectral annotation) will expand identification rates and structural insight from DDA datasets.

Conclusion
Methodical optimization of DDA parameters on the Agilent Revident LC/Q‑TOF—covering acquisition rates, precursor selection rules, exclusion strategies, and data storage settings—substantially improves MS/MS coverage and annotation potential in untargeted metabolomics. Directed MS/MS is effective for focused, biologically relevant lists, while iterative Auto MS/MS at high acquisition rates maximizes overall coverage at the cost of some loss in library match score. Choosing settings requires explicit trade‑offs tailored to chromatographic behavior, sample complexity, and the identification goals of the study.

Instrumentation used

• Agilent Revident LC/Q‑TOF with Agilent MassHunter Data Acquisition for TOF/Q‑TOF LC/MS v12.1 (and later).
• Agilent 1290 Infinity II Bio LC (compatible with Infinity III Bio LC).
• InfinityLab Poroshell 120 HILIC‑Z (2.1 × 150 mm, 2.7 µm).
• Agilent ZORBAX Eclipse Plus C18 column (2.1 × 50 mm, 1.8 µm) for plant extract tests.
• Agilent Captiva EMR cartridges for plasma extraction.
• MassHunter Qualitative Analysis v12.0/13.0 for library matching and scoring.

References

1. Sumner LW, Amberg A, Barrett D, et al. Proposed Minimum Reporting Standards for Chemical Analysis: Chemical Analysis Working Group (CAWG) Metabolomics Standards Initiative (MSI). Metabolomics. 2007;3(3):211–221.
2. Schymanski EL, Jeon J, Gulde R, et al. Identifying Small Molecules via High Resolution Mass Spectrometry: Communicating Confidence. Environ Sci Technol. 2014;48(4):2097–2098.
3. Durham SD, Yannel KE, Simmermaker C, Van de Bittner G, Bertram L, Cuthbertson D, Klein C. An End‑to‑End Untargeted LC/MS Workflow for Metabolomics and Lipidomics. Agilent Technologies application note, publication 5994‑8371EN, 2025.
4. Yannell KE, Durham S, Simmermaker C, Van de Bittner G. Uncovering More Biological Insights in Your Samples with Routine LC/Q‑TOF Workflows for Metabolites and Lipids. Agilent Technologies ASMS poster ThP‑085, 2024.
5. Zhao L, Juck M. Protein Precipitation for Biological Fluid Samples Using Agilent Captiva EMR–Lipid 96‑Well Plates. Agilent application note 5991‑9222EN, 2018.
6. Yannell KE, Simmermaker C, Van de Bittner G, Cuthbertson D. An End‑to‑End Targeted Metabolomics Workflow. Agilent application note 5994‑5628EN, 2023.
7. Ferrer I, Thurman EM, Zweigenbaum JA. Auto MS/MS and Identification of Unknowns in Water Samples. Agilent technical overview 5994‑0322EN, 2018.
8. Wu L, Wong DL. In‑depth Peptide Mapping with Iterative MS/MS Acquisition on the Agilent 6545XT AdvanceBio LC/Q‑TOF. Agilent application note 5991‑8633EN, 2020.
9. Koelmel J, Sartain M, Salcedo J, et al. Improving Coverage of the Plasma Lipidome Using Iterative MS/MS Data Acquisition Combined with Lipid Annotator Software and 6546 LC/Q‑TOF. Agilent application note 5994‑0775EN, 2020.