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 × 10(6), 4·0 × 10(6) and 8·0 × 10(6) 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(ß) ) of tetracyclines and increase sensitivity to betalactams was implemented. Significant effects of Log[S] and incubation time [It] on the CC(ß) of betalactams and tetracyclines were observed. Finally, high value of global desirability (D = 0·853), adequate betalactams CC(ß) (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(ß) (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.

Optimization of bioassay for tetracycline detection in milk by means of chemometric techniques.

AIMS: In this study, a microbiological method of dichotomous response using Bacillus cereus was designed and optimized to detect tetracyclines (TCs) at concentrations near to the maximum residue limits (MRLs). METHODS AND RESULTS: In a first stage, the response time of bioassay was reduced to 5 h when the logarithm of spore concentration (log S) was increased. Later, a Plackett Burman design (2(6-3)) was analysed using logistic regression model. This design indicates significant effects of log S and chloramphenicol (CAP) on the detection limit (DL) of TC. Then, the response surfaces (RS) of the TCs DTs as a function of log S and CAP were plotted using a Dohlert design and the logistic regression model. These RS show a linear decrease with the raise of CAP and a quadratic effect of log S. Finally, the DTs of TC (109 µg l(-1)) and oxytetracycline (100 µg l(-1)) were adjusted to their MRLs through the desirability function. CONCLUSIONS: By successive application of experimental design techniques could be optimized a bioassay for the detection of TC residues in milk. The best conditions have been achieved when the assay was made with log S = 5·12 and CAP = 470 µg l(-1). SIGNIFICANCE AND IMPACT OF THE STUDY: Experimental design techniques together with the logistic regression model and the desirability function represent an adequate tool for the optimization of a bioassay with binary response.

Robust experimental design for optimizing the microbial inhibitor test for penicillin detection in milk.

AIMS: To use experimental design techniques and a multiple logistic regression model to optimize a microbiological inhibition test with dichotomous response for the detection of Penicillin G in milk. METHODS AND RESULTS: A 2(3) x 2(2) robust experimental design with two replications was used. The effects of three control factors (V: culture medium volume, S: spore concentration of Geobacillus stearothermophilus, I: indicator concentration), two noise factors (Dt: diffusion time, Ip: incubation period) and their interactions were studied. The V, S, Dt, Ip factors and V x S, V x Ip, S x Ip interactions showed significant effects. CONCLUSIONS: The use of 100 microl culture medium volume, 2 x 10(5) spores ml(-1), 60 min diffusion time and 3 h incubation period is recommended. In these elaboration conditions, the penicillin detection limit was of 3.9 microg l(-1), similar to the maximum residue limit (MRL). Of the two noise factors studied, the incubation period can be controlled by means of the culture medium volume and spore concentration. SIGNIFICANCE AND IMPACT OF THE STUDY: We were able to optimize bioassays of dichotomous response using an experimental design and logistic regression model for the detection of residues at the level of MRL, aiding in the avoidance of health problems in the consumer.