Statistical evaluation of analytical curves for quantification of pesticides in bananas

Statistical evaluation to validate matrix-matched calibration for standardized beany odor compound quantitation in yellow pea flour using HS-SPME-GC-MS

“…As proposed by Gomes et al , a three-step scrutiny process of linearity was used and is as follows: examination of heteroskedasticity, visual inspection of the residual plot, and evaluation of outliers and regression parameters. 9…”

“…After the application of WLSR, the t r -test (eqn (7)) was used to verify the quality adjustment of the WLS regressed calibration curves. 9 A linear range is achieved and tested when t r w is no less than the corresponding critical value ( t crit ) at 95% level of confidence.where r w is the coefficient of determination obtained by WLSR and N x is the number of points used to establish the calibration curve.…”

“…However, like the sampling variation in YP flour and the batch effect of sample loading, the heteroskedasticity cannot be ignored and must be checked when applying regression analysis to establish calibration curves. 11 Under such conditions when heteroskedasticity is detected, 9 the assumption for OLSR is no longer validated since the error variances throughout the same concentrations of beany odor standards (replicated independent variables) are unequal. 12 Alternatively, common data pretreatment such as power transformation (the Box–Cox transformation, variance-stabilizing transformation), randomization tests, non-parametric models (the Kruskal–Wallis test), and weighted least squares regression (WLSR) can be introduced to minimize or correct the heteroskedasticity.…”

“…Thus, in this work, WLSR was adopted to directly balance and reweight the variances among observations and was implemented to establish the calibration curves. 9…”

“…However, applying a comparable deodorized matrix-matched flour that can best imitate the original YP flour is crucial for the ESC method, since ignoring the matrix effect leads to the loss of the figure of merit, such as linearity, limitation of detection/quantitation (LOD/LOQ), and precision. 9 Although a handful of studies have reported the quantitation of beany odor compounds in YP flour via interpolation from the ESC method, 3,10 the deodorized matrices used were completely different. Deodorized YP flour seemed to be the best candidate as it has the same constituents as the original YP flour, 3 although deodorized starch is the most affordable and commonly used one.…”