Scope and limitations of principal component analysis of high resolution LC-TOF-MS data: the analysis of the chlorogenic acid fraction in green coffee beans as a case study

Within this contribution we have analysed aqueous methanolic extracts by LC-ESI-TOF-MS of a total of 38 green bean coffee samples, which vary in terms of coffee variety and processing conditions. The LC-MS data have been analysed by principal component analysis (PCA) using different PCA processing parameters using an unsupervised non-targeted approach as well as a knowledge-based targeted approach. Furthermore, different normalisation and scaling algorithms have been applied to the PCA dataset. The scope and limitation of the various PCA parameters are discussed with respect to the ability to differentiate between samples of different groups, including different coffee varieties (Arabica or Robusta coffee) or different processing parameters and with respect to the information content of the PCA analysis on a molecular level. We could show that while distinction between different groups of samples can be successfully carried out independent of PCA parameters employed, identifying molecular markers rationalising differentiation between sample groups varies significantly between PCA parameters and requires careful choice as well as critical evaluation.