Determination of the detection limit using the four-parameter logistic model for the double-antibody sandwich ELISA for the rapid detection of Bacillus cereus in food

Abstract

Bacillus cereus is increasingly recognized as one of the major causes of food poisoning in the industrialized world. In a previous publication, a sandwich ELISA is developed for the detection of B. cereus whole cells. The curve showed a sigmoidal calibration curve, but the model used was linear. This study remodels the sigmoidal data obtained using a four-parameter logistic (4PL) equation that can cover the whole range of data obtained. The sigmoidal ELISA calibration curve was efficiently fit by the 4-PL equation. A good correlation coefficient value of 0.996 was obtained indicating good fitting. The LOD value was 5.36 x104 CFU/mL with the 95% confidence interval from 1.07x104 to 9.64x104 CFU/mL, which is about one order of magnitude less sensitive that the previously claimed LOD. The modelling exercise shows that the 4PL equation can fit the ELISA curve very well.