Chemometrics as an aid in authentication

Thousands of chemical compounds have been identified in oils and fats, although only a few hundred are used in authentication. This means that each object (food sample) may have a unique position in an abstract n-dimensional hyperspace. A concept that is difficult to interpret by analysts as a data matrix exceeding three features already poses a problem. The art of extracting chemically relevant information from data produced in chemical experiments by means of statistical and mathematical tools is called chemometrics. It is an indirect approach to the study of the effects of multivariate factors (or variables) and hidden patterns in complex sets of data. Chemometrics is routinely used for (a) exploring patterns of association in data, and (b) preparing and using multivariate classification models. The arrival of chemometrics techniques has allowed the quantitative as well as qualitative analysis of multivariate data and, in consequence, it has allowed the analysis and modelling of many different types of experiments. 

The general concept of authentication, on the other hand, is not only circumscribed to adulteration but also includes aspects such as characterization, geographical origin of foodstuffs, extraction and processing systems, among others. Thus, the combined effect of a great amount of data, supplied by the current sophisticated instrumentation and the various issues included inside the concept of authenticity means that the analyst faces a complex situation. We cannot therefore think that we only need to put data in one side of a black box (the computer) and get the results from the other side. 

The chemometric procedures that are currently applied in empirical investigations have been notably improved in recent years with the assistance of computer science. Researchers have passed from initial application of univariate analyses to extensive use of multivariate procedures in less than one decade. This qualitative step has been possible because, first, the new sophisticated analytical instruments are now able to analyse dozens of chemical compounds in hundreds of samples daily, and, second, due to personal computers that can work with a great diversity of software packages. 

The application of mathematical algorithms has allowed conclusions to be arrived at that were unthinkable only a few decades ago. However, the conclusions 

Exploratory data analysis (EDA). This analysis, also called pretreatment of data , is essential to avoid wrong or obvious conclusions. The EDA objective is to obtain the maximum useful information from each piece of chemico-physical data because the perception and experience of a researcher cannot be sufficient to single out all the significant information. This step comprises descriptive univariate statistical algorithms (e.g. mean, normality assumption, skewness, kurtosis, variance, coefficient of variation), detection of outliers, cleansing of data matrix, measures of the analytical method quality (e.g. precision, sensibility, robustness, uncertainty, traceability) (Eurachem, 1998) and the use of basic algorithms such as box-and-whisker, stem-and-leaf, etc. 

---

Andrea, T. A. Kalayeh, H. (1991). Applications of neural networks in quantitative structure-activity relationships of dihydrofolate reductase inhibitors. Journal of Medicinal Chemistry. Vol. 34, pp. 2824-2836. ISSN 0022-2623 Aparido, R. Aparicio-Ruiz, R. (2002). Chemometrics as an aid in authentication. In Oils and Fats Authentication, M Jee (Ed.), 156-180, Blackwell Publishing and CRC Press, ISBN 1841273309, Oxford, UK and FL, USA Bishop, CM. (2000). Neural Networks for Pattern Recognition, Oxford University Press, ISBN 0198538642, NY, USA... 

---