Decoding Dietary Signatures in Human Fecal Metabolomes Using a Novel timsTOF Platform

Decoding Dietary Signatures in Human Fecal Metabolomes: Summary of timsTOF-Based Deep Profiling of NIST RM 8048

Importance of the topic

Human fecal metabolomes encode combined signals from host metabolism, gut microbiota, and diet. Comprehensive and reproducible characterization of this complex matrix is essential for nutritional studies, biomarker discovery, microbiome research, and interlaboratory method harmonization. The availability of a pooled, well‑characterized reference material (NIST RM 8048) creates an opportunity to benchmark analytical workflows and to generate standardized digital metabolome archives supporting method development and comparison across platforms.

Objectives and study overview

The study aimed to generate an initial deep profiling benchmark for NIST RM 8048 by combining complementary chromatographic separations with trapped ion mobility spectrometry–time‑of‑flight mass spectrometry (TIMS‑QTOF). Specific goals were to:

• Maximize coverage of polar, semi‑polar, and non‑polar metabolites and lipids using multiple extraction and LC methods.
• Leverage multidimensional data (m/z, retention time, MS/MS, collisional cross section) to increase annotation confidence.
• Compare feature occurrence between fecal pools from vegetarian and omnivorous donors and identify diet‑associated markers.

Methodology

Sample preparation and extraction strategies were tailored to molecular polarity:

• HILIC extraction/analysis for polar metabolites (20% ACN / 80% H2O extraction; Waters BEH Z‑HILIC column).
• Reversed‑phase (RP) metabolomics for semi‑polar species (25% MeOH / 25% iPrOH / 50% H2O extraction; Phenomenex Kinetex C18 column).
• Lipidomics for non‑polar species (50% MeOH / 50% MTBE extraction; Waters Cortecs UPLC C18 column).

Chromatographic mobile phases and ionization modes were tuned to capture acidic and basic species in both positive and negative electrospray modes. Data acquisition used TIMS‑QTOF with PASEF/MoRE scan modes across m/z ranges appropriate for metabolites and lipids. Feature detection, recalibration and initial annotation were performed in MetaboScape 2026b; further annotation used R packages from the RforMassSpectrometry ecosystem and in‑silico tools (Sirius with CSI:FingerID). Lipid‑specific annotation employed rule‑based approaches and CCS matching/prediction (LipidCCS). Blank correction and replicate filtering (detected in at least 2/3 replicates per group and >1.5× blank) were applied to define robust features.

Used instrumentation

The study used a timsMetabo TIMS‑QTOF platform (Bruker) with the following LC hardware and software:

• Columns: Waters BEH Z‑HILIC (100 × 2.1 mm, 1.7 µm), Phenomenex Kinetex C18 (100 × 2.1 mm, 1.7 µm), Waters Cortecs UPLC C18 (150 × 2.1 mm, 1.6 µm).
• Mass spectrometer: TIMS‑QTOF (timsMetabo) operated with PASEF/MoRE acquisition modes for high‑speed MS/MS and mobility separation.
• Software/tools: MetaboScape 2026b for peak detection and initial annotation; R (RforMassSpectrometry packages) and Sirius with CSI:FingerID for structure elucidation; CCS prediction tools such as LipidCCS.

Note that Bruker instruments and software were used throughout the workflow.

Main results and discussion

Key findings from profiling NIST RM 8048 include:

• Detection of several thousand reproducible features across complementary LC methods and ionization modes. Lipidomics RP(+) yielded the largest feature set (~8,400 features before detailed filtering), while other methods detected several thousand each, indicating broad coverage from polar to non‑polar chemistry.
• High overlap between vegetarian and omnivorous pooled samples: roughly 70–80% of features were common to both donor groups, depending on method and polarity, while a subset of diet‑specific or diet‑enriched features was observed.
• Multidimensional annotation (m/z, MS/MS fragmentation, retention time, and measured CCS) increased confidence in putative identifications. Example: cholic acid and related bile acids were annotated using combined MS2, RT and CCS data matched against references.
• Discovery of putative novel species including fatty acid esters of 3‑hydroxy bile acids (e.g., a FA 16:1 ester of cholic acid or similar) supported by detection of multiple adducts and concordant in‑silico annotations, representing promising targets for follow‑up structural confirmation.
• Many features remain unannotated, especially within lipidomics, with some showing higher abundance in vegetarian samples — these represent priorities for targeted structural work and biological follow‑up.

The integration of TIMS (providing CCS) with high‑resolution MS/MS and orthogonal chromatographic separations proved effective at resolving isobaric species and prioritizing confident annotations in a chemically complex fecal matrix.

Benefits and practical applications of the method

The described workflow delivers several practical advantages:

• Comprehensive metabolome and lipidome coverage from a single reference material using a tiered extraction and LC approach.
• Improved annotation confidence via orthogonal data dimensions (m/z, RT, MS/MS, CCS), reducing false positives in complex matrices.
• Generation of digital metabolome archives that can serve as benchmarks for method development, instrument comparison, and interlaboratory harmonization.
• Identification of diet‑associated molecular signatures useful for nutritional epidemiology, microbiome‑diet interaction studies, and biomarker development.

These assets support both discovery research and the eventual development of targeted assays for clinical, nutritional, and environmental studies.

Future trends and potential applications

Anticipated developments and opportunities building on this work include:

• Systematic structural confirmation of putative novel compounds using orthogonal techniques (e.g., NMR, targeted MS/MS with authentic standards, chemical derivatization) to move from putative to confirmed IDs.
• Expansion of CCS libraries and public spectral repositories to accelerate cross‑platform annotation and automated identification.
• Application of machine learning and improved in‑silico prediction models to prioritize unknowns and predict biological origin (host vs microbial vs dietary).
• Use of standardized digital metabolome archives derived from reference materials like NIST RM 8048 for interlaboratory comparisons, QA/QC of clinical and research pipelines, and regulatory method validation.
• Translation of discovery signals into targeted quantitative assays for diet‑microbiome‑health studies and routine monitoring.

Conclusion

Combining complementary extraction chemistry and chromatographic separations with TIMS‑QTOF acquisition and multidimensional annotation yields a deep, high‑confidence metabolomic and lipidomic characterization of NIST RM 8048. The approach detects thousands of reproducible features, highlights diet‑associated differences, and reveals novel candidate molecules for further study. The resulting digital metabolome archive forms an initial benchmark for method development and harmonization in fecal metabolomics.

References

• NIST Reference Material 8048 (human fecal reference material) — used as the matrix for profiling.
• MetaboScape 2026b (Bruker) — peak detection, recalibration and initial annotation software used in this study.
• Sirius with CSI:FingerID — in‑silico structure elucidation and MS/MS annotation tool utilized for putative identifications.
• LipidCCS — CCS prediction/annotation resource applied to lipid identifications.
• Cruz et al. — previous study referenced for known metabolites detected in fecal samples (citation as mentioned in the source document).