Multivariate statistical methods for the assessment of the quality and the authenticity of saffron
Saffron is characterized by three main secondary metabolites: crocin, picrocrocin, and safranal.
These metabolites confer to saffron its quality.
Saffron is considered as the most expensive spice of the world due to its limited production and high demand.
Consequently, saffron was the subject of many ways of adulteration with different natural or synthetic substances to enhance its color and to increase its weight.
Therefore, authentication of saffron was always and still continues to be a challenging issue.
Currently, the quality of saffron is certified in the international trade market following the ISO 3632 normative by using analytical methods like spectrophotometry, chromatography, spectroscopy, and capillary electrophoresis.
This is called chemical fingerprinting.
Recently, molecular methods are used to determine adulteration in saffron.
This is called genetic or DNA fingerprinting.
Chemical and molecular fingerprinting are generating a huge amount of raw data that require sophisticated data analysis tools for extracting useful information.
The raw data are cleaned and transformed.
This constitutes the pre-treatment step which includes removing baseline artifacts, peak-picking, alignment and normalization, and scaling and transformation.
Then, the cleaned and transformed data are processed, either for calibration or classification, using different multivariate statistical methods like principal component analysis, hierarchical cluster analysis, partial least squares regression, partial least squares discriminant analysis, etc.
Finally, the developed statistical models should be validated to evaluate the reliability of their performances before using them, in practice, to authenticate saffron samples.
This research work will present the main analytical methods used for assessing the quality of saffron.
Then, the most used spectroscopic and chromatographic data pre-treatment tools will be discussed.
Finally, the most common multivariate statistical methods will be presented with how to validate them.
Douaik, A. (2017). Multivariate statistical methods for the assessment of the quality and the authenticity of saffron. Acta Hortic. 1184, 173-178
DOI: 10.17660/ActaHortic.2017.1184.25
https://doi.org/10.17660/ActaHortic.2017.1184.25
DOI: 10.17660/ActaHortic.2017.1184.25
https://doi.org/10.17660/ActaHortic.2017.1184.25
data modeling, model validation, preprocessing, supervised, unsupervised
English