Implementing Tile-Based Fisher Ratio Analysis of GC×GC-TOFMS Data to Obtain a Master Peak Table of All Detected Analyte Compounds in Many Petroleum-Based Samples

Importance of Topic

High‐resolution analysis of complex petroleum‐based mixtures is critical for quality control, regulatory compliance, and forensic investigations. Comprehensive two‐dimensional gas chromatography coupled with time‐of‐flight mass spectrometry (GC×GC‐TOFMS) generates vast data sets with superior separation power, but poses challenges in automated peak detection and alignment across multiple samples. Implementing advanced chemometric approaches improves the generation of consistent, high‐fidelity peak tables for comparative studies and fingerprinting.

Objectives and Study Overview

This collaborative study between the University of Washington and Chevron Technical Center aimed to develop a robust, largely automated workflow to extract a master peak table of all analytes detected across numerous petroleum fuel samples. The goals were to leverage a tile‐based Fisher ratio (F‐ratio) method for class discrimination, incorporate blank‐based p‐testing to confirm detectability, and assemble a reproducible peak list that captures trace and major components under varying sample classes.

Methodology

A tiling approach divides GC×GC chromatograms into discrete regions and calculates an F‐ratio per tile to identify areas with statistically significant concentration differences between predefined sample classes. Major steps included:

• Generation of chromatographic tiles using ChromaTOF Tile module.
• Calculation of between‐class versus within‐class variance to rank tiles by F‐ratio.
• Aggregation of highest‐scoring tiles into a preliminary hit list.
• Application of blank chromatogram replicates to derive 99% confidence interval limits per tile.
• Binomial p‐testing across three blank replicates to confirm true signal detectability in each tile.
• Construction of a Master Peak Table (MPT) by combining F‐ratio hits that pass blank‐based detectability criteria.

Used Instrumentation

The analytical platform comprised GC×GC‐TOFMS with the following components:

• First‐dimension column: Rxi-1 (30 m × 0.25 mm i.d. × 0.25 µm film).
• Second‐dimension column: Rxi-17MS (2.0 m × 0.18 mm i.d. × 0.20 µm film).
• Thermal modulator for periodic injection into the second dimension.
• Time‐of‐flight mass spectrometer providing full‐scan detection.
• ChromaTOF software for baseline processing, tile generation, and F‐ratio calculation.

Main Results and Discussion

The workflow was applied to nine distinct fuel classes, capturing both major compounds eluting early (within 15–27 minutes) and late‐eluting trace constituents. Key findings:

• A total of 719 unique tile hits passed both F‐ratio ranking and blank detectability criteria.
• F‐ratio values spanned several orders of magnitude, reflecting variability in analyte abundance across classes.
• Reproducibility was high: only 6.3% of peak assignments varied across replicates, indicating robust peak selection.
• Identifications included aromatic species (e.g., 1‐methylnaphthalene), branched alkanes, and cyclopentene derivatives, consistent with expected fuel compositions.

Benefits and Practical Applications

This tile‐based Fisher ratio method offers multiple advantages for petroleum analysis:

• Reduction of retention time shift effects due to tiled segmentation.
• Enhanced recovery of low‐abundance analytes with statistical confidence.
• Minimized user intervention for peak picking and alignment.
• Applicability to diverse sample matrices beyond fuels for fingerprinting and QA/QC.

Future Trends and Applications

Opportunities to extend this approach include:

• Exploring alternative chemometric metrics (e.g., variance ranking, supervised learning scores) tailored to experimental designs.
• Integration with machine‐learning algorithms for automated compound class prediction.
• Application to environmental, food, and clinical metabolomic studies requiring comprehensive profiling.
• Development of open‐source workflows to democratize high‐dimensional data analysis.

Conclusion

The tile‐based Fisher ratio workflow combined with blank‐based p‐testing delivers a powerful, reproducible method for generating a comprehensive master peak table from GC×GC‐TOFMS data. This automated strategy reduces retention time alignment burdens and reliably captures both major and trace analytes across multiple sample classes, facilitating comparative studies and quality assessments in petroleum analytics.

References

R. Halvorsen, W. Ma, C. Cain, H. Ingham, R. Mohler, R. Synovec, "Implementing Tile‐Based Fisher Ratio Analysis of GC×GC‐TOFMS Data to Obtain a Master Peak Table of All Detected Analyte Compounds in Many Petroleum‐Based Samples," Journal of Chromatography Open 8 (2025) 100249.