Food Research Institute

POLOVKA, M. – SUHAJ, M.
Classification and prediction of gamma-irradiation of ten commercial herbs and spices by multivariate evaluation of properties of their extracts
Journal of Food and Nutrition Research, 52, 2013, No. 1, s. 45-60

Martin Polovka, Department of Chemistry and Food Analysis, VUP Food Research Institute, Priemyselná 4, P.?O.?Box 25, SK 82475 Bratislava, SlovakiTel.: +421 2 502 37 195, fax: +421 2 5557 1417, e-mail: polovka@vup.sk

Summary: Antioxidant activity, dry matter content, extractability, total phenolic compounds content and CIE L*a*b* colour characteristics of methanolic extracts of 10 spices exposed to gamma-irradiation doses from 0 kGy (reference) to 30 kGy were evaluated in different time intervals after the irradiation by means of UV-VIS spectrophotometry in order to assess the influence of gamma-irradiation and post-irradiation storage on spices quality. Experimental data revealed that gamma-irradiation itself did not cause so dramatic changes as the subsequent post-irradiation storage, or, that the changes were, at least, comparable. Multivariate statistical methods (factor analysis, canonical discriminant analysis and kth-nearest neighbour classification) facilitated differentiation of irradiated spices from the corresponding references and gamma-radiation dose prediction on the basis of processing of experimentally determined characteristics of their extracts. In case of samples exposed to 0 kGy, 10 kGy and 30 kGy, differentiation and classification correctness higher than 85% was reached. In case of 0 kGy and 30 kGy samples differentiation, for 8 spices absolutely correct recognition was achieved and, for remaining 2 spices, still sufficient recognition correctness of 75% and 92.8%, respectively, was achieved. By the statistical models used, more than 90% correctness of dose prediction was achieved in the majority of cases.

Keywords: spices; gamma-irradiation; antioxidant properties; dose estimation; discrimination; classification

Download:
  (pdf, 259.51 Kb, 2368x)