Elsevier

Analytica Chimica Acta

Volume 936, 14 September 2016, Pages 245-258
Analytica Chimica Acta

Combined untargeted and targeted fingerprinting with comprehensive two-dimensional chromatography for volatiles and ripening indicators in olive oil

https://doi.org/10.1016/j.aca.2016.07.005Get rights and content

Highlights

  • Fingerprinting of Extra Virgin Olive Oil volatiles by comprehensive two-dimensional gas chromatography mass spectrometry.

  • Combined strategy for the most inclusive and effective fingerprinting of complex volatile fraction.

  • Integrated and automated data processing for untargeted and targeted fingerprinting by peak-regions and visual features.

  • Reliable and consistent retrospective analysis on 2D patterns and highly informative pair-wise comparison.

Abstract

Comprehensive two-dimensional gas chromatography (GC × GC) is the most effective multidimensional separation technique for in-depth investigations of complex samples of volatiles (VOC) in food. However, each analytical run produces dense, multi-dimensional data, so elaboration and interpretation of chemical information is challenging.

This study exploits recent advances of GC × GC-MS chromatographic fingerprinting to study VOCs distributions from Extra Virgin Olive Oil (EVOO) samples of a single botanical origin (Picual), cultivated in well-defined plots in Granada (Spain), and harvested at different maturation stages. A new integrated work-flow, fully supported by dedicated and automated software tools, combines untargeted and targeted (UT) approaches based on peak-region features to achieve the most inclusive fingerprinting.

Combined results from untargeted and targeted methods are consistent, reliable, and informative on discriminant features (analytes) correlated with optimal ripening of olive fruits and sensory quality of EVOOs. The great flexibility of the UT fingerprinting here adopted enables retrospective analysis with great confidence and provides data to validate the transferability of ripening indicators ((Z)-3-hexenal, (Z)-2-hexenal, (E)-2-pentenal, nonanal, 6-methyl-5-hepten-2-one, octane) to external samples sets. Direct image comparison, based on visual features, also is investigated for quick and effective pair-wise investigations. Its implementation with reliable metadata generated by UT fingerprinting confirms the maturity of 2D data elaboration tools and makes advanced image processing a real perspective.

Introduction

Comprehensive two-dimensional gas chromatography (GC × GC) is the most effective multidimensional separation technique for in-depth investigations of complex samples of volatiles in food [1]. The combination, in a single analytical platform, of two separation dimensions with mass spectrometric detection and, when possible, automated sample preparation, delivers highly efficient sample profiling (detailed analysis of single molecular entities) and fingerprinting (rapid, high-throughput screening of samples for distinctive analytical signatures) [2].

Each analytical run produces dense, multi-dimensional data, so elaboration and interpretation of chemical information is a challenging task. In addition, food samples generally have a high-degree of chemical multidimensionality [3] thus creating highly complex analytical challenges. In this context, data elaboration strategies should implement smart and productive processes, preferably with a high degree of automation, to make cross-samples analysis efficient and informative.

Within the existing methodologies for GC × GC data elaboration [4], [5], the approach based on peak-region features has been very effective because of its comprehensive and uniform treatment of information from each sample constituent, both knowns and unknowns. Each single chemical entity is characterized by its chromatographic and spectrometric parameters (retention time in both dimensions, detector response, and mass spectral information) and by its absolute and relative position within the pattern of all detectable constituents. As a consequence, the 2D peak-retention pattern of a sample is a diagnostic fingerprint, informative of its composition; and pattern recognition approaches can be successfully applied to improve effectiveness and productivity in multi-sample data elaboration.

Although these concepts are not new for the GC × GC community [6], the full automation of these procedures and their implementation in commercial software packages has been achieved only recently. This has limited both routine adoption of the technique for food analysis and investigative strategies for profiling [2], [7].

Analysis of olive oil volatiles is a challenging and important problem and GC × GC can yield deeper knowledge of the composition of this fraction offering new perspectives for quality and authenticity assessment [8].

In spite of the great potential of GC × GC, few studies are available in this field. Vaz Freire et al. [9]first proposed an image-features approach, or more generally a pattern recognition methodology, to investigate the characteristic distribution of volatiles from oils. They adopted open-source image analysis software (Image J, National Institutes of Health) to extract information from small 2D regions located over the separation space and, by Principal Component Analysis (PCA), selected those regions with the highest discrimination potential. Then, they used targeted profiling to locate known analytes within informative 2D regions.

In 2010, Cajka and co-workers [10] exploited the targeted profiling potential of GC × GC-ToF-MS and identified 44 analytes able to discriminate samples of different geographical origin and production year. More recently, Purcaro et al. [8] combined targeted and untargeted analysis with the goal of a chemical blueprint of olive oil aroma defects. This inter-laboratory study confirmed the reliability of GC × GC for detailed profiling of olive oil volatile fractions and introduced an iterative strategy [11], [12] to locate sensory-relevant analytes efficiently.

This study exploits the most recent advances of GC × GC-MS chromatographic fingerprinting to study VOC distributions from Extra Virgin Olive Oil (EVOO) samples of a single botanical origin (Picual), cultivated in well-defined plots in a single region (Granada, Spain), and harvested at different maturation stages. The principal interest in this application is the quality characteristics related to optimal ripening of olive fruits [13], [14], [15], [16], [17], [18], [19], [20], [21] and, as a consequence, olive oil classification and perceivable sensory quality [22], [23]. In particular, this study proposes an integrated work-flow, fully supported by dedicated software tools, that performs cross-samples comparisons by contemporarily considering characteristic distributions (i.e., sample fingerprints) of both known and unknown compounds. This work-flow integrates both untargeted and targeted (UT) fingerprinting to realize the most comprehensive results, and so is termed UT fingerprinting. Challenges of retrospective analysis and immediacy of image fingerprinting also are discussed because of the advantages they offer in specific investigations.

Section snippets

Reference compounds and solvents

Pure reference standards of α-thujone, used as Internal Standard (ISTD), at a concentration of 100 mg/L in dibuthyl phthalate, and n-alkanes (n-C9 to n-C25), used for linear retention index (ITS) determination, at a concentration of 100 mg/L in cyclohexane, were supplied by Sigma-Aldrich (Milan, Italy).

Solvents for n-alkanes dilution (toluene and cyclohexane HPLC-grade) and dibuthyl phthalate also were from Sigma-Aldrich.

Olive oil samples

Olive oil samples of Picual variety, harvested in 2014, were supplied by

Results and discussion

The goal of this study was to evaluate the potential of combining Untargeted and Targeted 2D data elaboration approaches based on untargeted peak-region features, target peaks, and visual features to approach to the most inclusive fingerprinting within EVOO volatiles: the UT fingerprinting strategy. Based on a sampling design focused on a single botanical variety and well-defined geographical locations, VOCs fingerprints were interpreted as a function of ripening stage and oil quality.

This

Conclusions

This study evidences and emphasizes the potentials of fingerprinting based on GC × GC-MS separations and highlights the synergism between untargeted and targeted methodologies to investigate complex fractions of volatiles in depth. Their combination enables to achieve the most inclusive/comprehensive fingerprinting (UT fingerprinting) and if compared to previous studies, the degree of automation implemented in the data elaboration work-flow is promising. Experimental results on EVOO volatiles

Acknowledgments

The authors are grateful to “GDR Altiplano de Granada” (Spain) for olives samples.

The research was carried out thanks to the support “International mobility program for young researchers (PhD)” by University of Granada and CEI BioTic Granada.

Note: S. E. Reichenbach has a financial interest in GC Image, LLC.

References (48)

  • D. Bressanello et al.

    Urinary metabolic fingerprinting of mice with diet-induced metabolic derangements by parallel dual secondary column-dual detection two-dimensional comprehensive gas chromatography

    J. Chromatogr. A

    (2014)
  • S.E. Reichenbach et al.

    Informatics for cross-sample analysis with comprehensive two-dimensional gas chromatography and high-resolution mass spectrometry (GCxGC-HRMS)

    Talanta

    (2011)
  • B. Hollingsworth et al.

    Comparative visualization for comprehensive two-dimensional gas chromatography

    J. Chromatogr. A

    (2006)
  • I. Romero et al.

    Validation of SPME-GCMS method for the analysis of virgin olive oils volatiles responsible for sensory defects

    Talanta

    (2015)
  • G. Luna et al.

    Characterisation of 39 varietal virgin olive oils by their volatile composition

    Food Chem.

    (2006)
  • M.T. Morales et al.

    Comparative study of virgin olive oil sensory defects

    Food Chem.

    (2005)
  • A. Raffo et al.

    Combined effects of reduced malaxation oxygen levels and storage time on extra-virgin olive oil volatiles investigated by a novel chemometric approach

    Food Chem.

    (2015)
  • C. Cordero et al.

    Comprehensive two-dimensional gas chromatography and food sensory properties: potential and challenges

    Anal. Bioanal. Chem.

    (2015)
  • C. Giddings

    Sample dimensionality: a predictor of order-disorder in component peak distribution in multidimensional separation

    J. Chromatogr. A

    (1995)
  • P.Q. Tranchida et al.

    Impact of comprehensive two-dimensional gas chromatography with mass spectrometry on food analysis

    J. Sep. Sci.

    (2016)
  • G. Purcaro et al.

    Toward a definition of blueprint of virgin olive oil by comprehensive two-dimensional gas chromatography

    J. Chromatogr. A

    (2015)
  • C. Cordero et al.

    Targeted and non-targeted approaches for complex natural sample profiling by GC×GC -qMS

    J. Chromatogr. Sci.

    (2010)
  • M.N. Franco et al.

    Influence of the fruit's ripeness on virgin olive oil quality

    J. Oleo Sci.

    (2015)
  • L. Martínez Nieto et al.

    Changes in phenolic compounds and rancimat stability of olive oils from varieties of olives at stages of ripeness

    J. Sci. Food Agric.

    (2010)
  • Cited by (0)

    1

    Federico Magagna and Lucia Valverde-Som, listed in alphabetical order, equally contributed to this work.

    View full text