Elsevier

Journal of Chromatography A

Volume 1218, Issue 3, 21 January 2011, Pages 518-523
Journal of Chromatography A

Recognition of volatile compounds as markers in geographical discrimination of Spanish extra virgin olive oils by chemometric analysis of non-specific chromatography volatile profiles

https://doi.org/10.1016/j.chroma.2010.11.045Get rights and content

Abstract

Chromatographic profiles obtained by headspace solid-phase microextraction (HS-SPME) coupled with gas chromatography (GC) were processed as continuous and non-specific signals through multivariate analysis techniques in order to select and identify the most discriminate volatile marker compounds related to the geographical origin of extra virgin olive oils. The blind analysis of the chromatographic profiles was carried out on several steps including preliminary mathematical treatments, explorative analysis, feature selection and classification. The results obtained through the application of stepwise linear discriminant analysis (SLDA) method revealed a perfect discrimination between the different Spanish geographical regions considered (La Rioja, Andalusia and Catalonia). The assignment success rate was 100% in both classification and prediction by using cross validation procedure. In addition, it must be noted that the proposed strategy was able to verify the geographical origin of the samples involving only a reduced number of discriminate retention times selected by the stepwise procedure. This fact emphasizes the quality of the accurate results obtained and encourages the feasibility of similar procedures in olive oil quality and traceability studies. Finally, volatile compounds corresponding to the predictors retained were identified by gas chromatography–mass spectrometry (GC–MS) for a chemical interpretation of their importance in quality virgin olive oils.

Introduction

Extra virgin olive oil is a valuable and prized foodstuff in Mediterranean countries. In order to guarantee and promote high quality extra virgin olive oils that show particular quality properties attributed to their geographical area of production, the European Union (EU) legislates and regulates the creation of protected designations of origin (PDO) [1], [2]. Oils labelled with this designation acquire an “added-value” which results in a higher price in the market. This fact highlights the importance of ensuring the authentication of these products with the aim of providing real protection for the rights of consumers and reliable producers from fraudulent practices involving, for instance, mixing with cheaper oils in order to make a profit. Therefore, food safety and traceability are essential challenges that must be faced in quality control tasks in order to achieve an objective differentiation among the different olive oils to ensure the declared origin and the quality labelled. According to these goals, reliable methods based on the evaluation of quality factors and properties are required.

In this sense, flavour is a remarkable factor that plays an important role in the determination of extra virgin olive oils authenticity. Distinctive sensorial properties can be attributed to a complex mixture of more than one hundred volatile compounds [3], [4] from different chemical classes, mainly aldehydes, alcohols, esters, hydrocarbons and ketones. In high quality virgin olive oils, C6 and C5 compounds enzymatically generated through the lipoxygenase (LOX) pathway represent the most important fraction of volatile profiles. [5], [6], [7].

The official procedure established by European legislation and the International Olive Oil Council (IOOC) [8], [9], [10] for the sensory analysis of olive oil's flavour involves notable problems related to poor repeatability and subjectivity [11]. Therefore, since the volatile profile of olive oils represents a fingerprint of the samples, it is reasonable to assume that more precise alternative strategies based on analysis of the volatile compounds are necessary in order to take advantage of the maximum amount of information present in the volatile fraction, to develop efficient classification approaches for evaluating olive oil authenticity.

A number of previous studies have reported the effectiveness of different analytical procedures in the development of strategies focusing on olive oil characterisation purposes [12], [13], [14], [15], [16], [17]. In this context, gas chromatography (GC) coupled with mass spectrometry (MS) has been the most widely employed, involving different extraction and pre-concentration techniques of the volatile fraction in order to extract minor compounds and obtain representative profiles [18], [19], [20]. The results reported showed that solid phase microextraction (SPME) technique [21], [22], had important advantages over other extraction procedures in the study of virgin olive oils. SPME is a sensitive, solvent-free, cost-efficient and fast method which allows the extraction and the concentration steps to be performed simultaneously. SPME in combination with GC–MS analysis has been successfully applied in olive oil characterisation and classification research to evaluate the effect of geographical origin, environmental conditions, olive varieties and detection of adulterations [23], [24], [25], [26], [27], [28], [29], [30], [31]. This analytical methodology has been traditionally associated with target analysis involving a first step of identification and quantification of specific markers existing in the volatile fraction. However, finding adequate and sufficiently specific markers is not always a quick and reliable procedure and therefore the quality assurance study may be tackled by considering the chromatographic profiles as continuous and non-specific signals, avoiding a priori identification of compounds. In this sense, realisable classification strategies involving foodstuff fingerprints have been developed in recent studies by applying different chemometrical tools mainly over spectroscopic and spectrometric matrices directly acquired on untreated food product samples [32], [33], [34], [35], [36], [37]. In the same way, chromatographic fingerprints have been studied in fields such as the petroleum industry and herbal medicine [38], [39], [40]. On the basis of these results it is reasonable to assume that multivariate analysis of chromatographic profiles may be an efficient strategy to recognize the volatile compound markers associated with the geographical discrimination of virgin olive oils.

The aim of the present study lies in identifying a series of volatile compounds able to discriminate extra virgin olive oils in accordance with geographical origin through the development of a strategy based on the blind analysis of the volatile fraction by means of combining the non-specific volatile profiles with multivariate analysis, in order to work directly with the whole fingerprint. The methodology proposed was based on the combined use of headspace (HS) SPME/GC–MS with the subsequent analysis of the chemical fingerprints using a SLDA method [41], [42] and a PCA data compression strategy to classify olive oils from different geographic origins. Finally, the identification of the volatile compounds selected for the geographical differentiation of the categories to be studied was carried out.

Therefore, the main novelty of the present work concerns the study of the potential of using non-specific volatile profiles to identify the volatile marker compounds related to the geographical discrimination of EVOOs through the development of a discrimination strategy based on combining the non-specific volatiles profiles with multivariate analysis. This procedure could be of great interest in olive oil authentication research because it involves the application of a large potential analytical technique in volatile component analysis (SPME/GC–MS) and presents a distinctive feature with respect to other previous studies carried out in this field because it avoids the need to carry out an a priori peak identification allowing direct work on the whole fingerprint. The proposed procedure was applied to the discrimination of extra virgin olive oils produced and manufactured in La Rioja, a region located in the north of Spain and distinguished by the certificated POD “Aceite de La Rioja” since 2004, from other protected extra virgin olive oils produced in southern and eastern regions of Spain according to their origin.

To our knowledge the identification of features related to the geographical origin of extra virgin olive oils through the blind analysis of chromatographic profiles has not been previously reported.

Section snippets

Extra virgin olive oil samples

The data set comprised a total of forty extra virgin olive oil samples from several Spanish regions, all of them labelled with the PDO quality trademark and collected in the same harvest. Nine of the samples corresponded to extra virgin olive oils from PDO “Aceite de La Rioja”, twenty-three were produced and manufactured in several PDOs from Andalusia, and eight samples came from PDO Siurana and PDO Les Garrigues (Catalonia). The samples were divided into three categories (Andalusia, La Rioja

Exploratory analysis

Since the proposed strategy is based on blind multivariate analysis of chromatographic profiles, areas of chromatographic peaks were not integrated; nor was a priori assignment of the peaks to its corresponding volatile compound carried out. The aligned and row profile scaled data set comprising 722 variables and 40 samples constituted the starting point for the pattern recognition analysis. The number of categories was defined according to the aim of discriminate protected extra virgin olive

Conclusions

In the present study, a classification methodology based on the blind analysis of non-specific chromatographic profiles was proposed in order to select and identify the most discriminate volatile markers associated with the geographical origin of extra virgin olive oils. The results reported showed that the combination of HS-SPME/GC analysis with SLDA method resulted in a satisfactory discrimination between the studied categories involving only six input variables. The subsequent identification

Acknowledgements

The authors wish to thank the Spanish Government (Ministerio de Ciencia e Innovación, Research Grant FPU and project ref. CTQ2008-03493) and the Local Government of La Rioja (Consejería de Educación, Cultura y Deporte, project ref. Fomenta 2007/06) for their financial support, as well as Professor Michele Forina for providing the Parvus package.

References (56)

  • M.T. Morales et al.

    J. Chromatogr. A

    (1994)
  • S. Vichi et al.

    J. Chromatogr. A

    (2003)
  • H.T. Lawless

    Food Qual. Prefer.

    (1999)
  • S. Mildner-Szkudlarz et al.

    Food Chem.

    (2008)
  • M.S. Cosio et al.

    Anal. Chim. Acta

    (2006)
  • A. Cimato et al.

    Sens. Actuators B

    (2006)
  • M.T. Morales et al.

    Food Chem.

    (2005)
  • C. Cerrato-Oliveros et al.

    J. Chromatogr. A

    (2005)
  • P. Zunin et al.

    J. Chromatogr. A

    (2004)
  • S. Vichi et al.

    Food Chem.

    (2007)
  • H. Lord et al.

    J. Chromatogr. A

    (2000)
  • T. Cajka et al.

    Food Chem.

    (2010)
  • M. Issaoui et al.

    Food Chem.

    (2010)
  • S. Mildner-Szkudlarz et al.

    Food Chem.

    (2008)
  • B. Baccouri et al.

    Food Chem.

    (2007)
  • G. Luna et al.

    Food Chem.

    (2006)
  • D. Tura et al.

    Food Chem.

    (2004)
  • M.J. Lerma-García et al.

    Food Chem.

    (2010)
  • M. Casale et al.

    Food Chem.

    (2010)
  • S. Lopez-Feria et al.

    Talanta

    (2008)
  • I. Esteban-Diez et al.

    Talanta

    (2007)
  • C. Xu et al.

    J. Chromatogr. A

    (2006)
  • A. Runcio et al.

    Food Chem.

    (2008)
  • M. Daszykowski et al.

    J. Chromatogr. A

    (2007)
  • A.M. van Nederkassel et al.

    J. Chromatogr. A

    (2006)
  • M. D’Imperio et al.

    Food Chem.

    (2007)
  • M. Casale et al.

    Talanta

    (2010)
  • F. Angerosa et al.

    Food Chem.

    (2000)
  • Cited by (0)

    View full text