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 (CC_β) of tetracyclines and increase sensitivity to betalactams was implemented. Significant effects of Log[S] and incubation time [It] on the CC_β of betalactams and tetracyclines were observed. Finally, high value of global desirability (D = 0·853), adequate betalactams CC_β (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 CC_β (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.