LIPID TUMOUR PROFILE

Abstract

We describe a method of generating a classification model capable of distinguishing between two biological states, the method comprising the steps of: (a) providing a training dataset, X, comprising concentrations of a plurality of lipids in a biological sample in a first state and a biological sample in a second state; (b) subjecting the training dataset X to Principal Components Analysis (PCA), in which the PCA analysis generates a transformation matrix, C, and a transformed dataset, Y.sub.l; (c) subjecting the transformed dataset Y.sub.l to Support Vector Machines (SVM) analysis, in which the SVM analysis generates a SVM model, S; (d) forming a classification model comprising (i) the transformation matrix C, and (ii) the corresponding SVM model S.