Validación estadística de la presencia en plantas de quimiotipos caracterizados por la concentración de componentes volátiles obtenida mediante GC-MS

Abstract

The name “chemotypes” indicates a group of individuals of a species which can be significantly distinguished from the rest from their chemical composition. Whereas the chemical polymorphism of several species is well known, the statistical significance of their differences in composition has been scarcely studied. The application of the k-means algorithm as a clustering technique, of the Mahalanobis distance Dmah as a measure of group separation and its comparison with Dmah values calculated from simulated data having a normal distribution, is proposed in order to assign a significance level to a single compound chemotype. The method is applied to volatile concentration data from samples of Thymus zygis subsp. zygis, Thymus zygis subsp. sylvestris and Lavandula luisieri.