A method for stepwise selection of peaks in NMR spectra from multiple groups is described. This method is based on initial peak-finding among the spectra and uses jacknife classi.cation performance as the basis for selection of peaks. The selection process is followed by the construction of correlation maps to identify sets of multiplets that are related to each of the selected peaks, aiding in the identi.cation of metabolites that are responsible for differences among the groups. For illustrative purposes, this methodology is applied to a data set that contains 52 spectra from renal cell carcinoma and normal renal tissue samples. The new method is denoted as StePSIM, Stepwise Peak Selection and Identification of Metabolites.
- Partial least squares
- Stepwise selection
ASJC Scopus subject areas
- Endocrinology, Diabetes and Metabolism
- Clinical Biochemistry