Application of Fourier transform to preprocessing chromatographic fingerprints of traditional Chinese medicine

• Using Fourier transform to overcome retention time shift in HPLC fingerprint of TCM.
• The influence of retention time shift on similarity estimation and PCA is eliminated.

Retention time shift is a significant impediment against analyzing chromatographic fingerprints of Traditional Chinese Medicine (TCM) with similarity estimation and pattern recognition methods. This work aims at developing a fingerprint preprocessing method that can eliminate the negative effect of retention time shift on similarity estimation and pattern recognition analysis in the quality evaluation of TCM. Both simulated chromatograms and experimental chromatographic fingerprints of Radix Puerariae thomsonii (RPT) were investigated in order to evaluate the practicality of the proposed method. Fourier transform (FT) was used to process these chromatographic signals. The magnitude data of these signals were calculated. In the obtained magnitude data, the negative effect of retention time shift was successfully excluded. Thus, the quality of the investigated RPT samples can be correctly evaluated by using the magnitude data as the input variables of similarity estimation and pattern recognition analysis. It is demonstrated that FT is a practicable and promising method for preprocessing chromatographic fingerprints of TCM prior to similarity estimation and pattern recognition analysis.