Use chemometric techniques in the optimization of a specific bioassay for betalactams in milk

Aims: A microbiological bioassay using Geoacillus stearothermophilus was optimized to detect betalactams at concentrations near to the Maximum Residue Limits (MRLs), with low cross‐specificity for tetracycline. Methods and Results: A factorial design (3 × 4) was used to evaluate the effects of concentration of spores (2·0 × 106, 4·0 × 106 and 8·0 × 106 spores ml−1) and incubation time (3·0, 3·5, 4·0 and 4·5 h) on the response of the bioassay. Then, desirability function to raise the detection capabilities (CCβ) of tetracyclines and increase sensitivity to betalactams was implemented. Significant effects of Log[S] and incubation time [It] on the CCβ of betalactams and tetracyclines were observed. Finally, high value of global desirability (D = 0·853), adequate betalactams CCβ (3·8 μg l−1 of penicillin ‘G’, 27 μg l−1 of oxacillin, 8·1 μg l−1 of ampicillin, 48 μg l−1 of cloxacillin) and high tetracyclines CCβ (5260 μg l−1 chlortetracycline, 1550 μg l−1 of oxytetracycline, 1070 μg l−1 of tetracycline) were calculated. Conclusions: The application of chemometric tools allows the optimization of a bioassay that detects betalactam residues in milk. The more robust conditions have been achieved in Log[S] = 6·30 and [It] = 4·20 h. Significance and Impact of the Study: The logistic regression model and the desirability function are adequate chemometric techniques to improve the properties of the methods, because it is possible to increase sensitivity and decrease cross‐specificity simultaneously.