Correlation between chromatographic fingerprint and antioxidant activity of Turnera diffusa (Damiana)

In the present contribution, the partial least squares (PLS) method was used to establish a correlation between the antioxidant activity (obtained by DPPH assay) and chromatographic profiles of TURNERA DIFFUSA extracts. Chromatograms were obtained using HPLC-DAD. A model was constructed using 40 samples with 2550 X variables corresponding to the responses obtained at different times; the Y variables consisted of experimental values of antioxidant activity of each extract (measured as EC₅₀). Prior to this analysis, alignment of chromatograms was performed based on consideration of seven high-intensity signals present in all samples. The PLS1 model was validated by cross-validations; its capacity was evaluated using correlation parameters R², root mean square error of calibration (RMSEC), and root mean square error of prediction (RMSEP). The best results were achieved with zero order chromatograms using five-point smoothing (R² = 0.96, RMSEC = 3.31, and RMSEP = 7.86). Under these conditions, the optimal number of components was five. The model was applied to the prediction of antioxidant activity of commercial products; no significant differences were found between the experimental and predicted antioxidant activities for 83 % of them.