Last PCR developments for the species specific detection of MBM in feed: transferability and single particle analysis

L2-3

Presently several PCR tests do exist to detect processed animal proteins (PAP) within feed by looking for DNA targets of a defined taxon (that will generally be at species level but it might be groups of species like ruminants for instance). These tests are in essence qualitative because the abundance of the analyte targeted by PCR is widely dependent on the sterilization process used during the PAP production. Even if the PCR-test can quantify the target looked for, it is of no interest for PAP-quantification because there is no straight relationship of these copy numbers with the quantity of PAP from which they are issued. It is nevertheless possible to design qualitative PCR tests with a limit of detection expressed in terms of minimal mass fractions of PAP that are detectable with a sufficient high probability just by considering what are the products generally present on the market and taking the more difficult ones for defining the LOD. Comparisons with blind samples have shown that several PCR tests are able to detect 0.1% of PAP issued from the animal taxons considered by the assay.

Up to now all the designed tests were used in the laboratory of their respective developer. Extending their use to other laboratories implies to address the transferability challenge of how to express the cut-off value of these tests. Practically within a laboratory this parameter is expressed in Ct but this parameter is merely relative and cannot be exchanged between laboratories because it is highly dependent for instance on the type of thermocycler used or on the mastermix involved in the reaction.

On the first thermocycler on which the assay were developed at CRA-W, we came empirically to a cut-off off value of 40 cycles after analyzing a lot of samples and especially negative ones that were considered as such to the best of our knowledge and checked by use of several other techniques than PCR. The question raised is then what would this value of cut-off become if used on another thermocycler or with another mastermix. We thought about expressing the cut-off in terms of copy numbers of the target as this is an absolute figure. However it appeared that a cut-off value of 40 corresponded to less than 1 copy of the target. We therefore designed a statistical approach to cope with that paradigm. The cut-off value is then defined as being the cycle number for which one copy of the target is detected with a probability of 95%. The protocol that was set up to measure this cut-off has been submitted to an interlaboratory trial to check its fitness for purpose. Adapted samples had to be chosen to really test the discrimination capacity of the test. Effects linked to the stochastic distribution of copy number of targets that are inherent to pipeting had to be circumvented by appropriate replicates. The test involved 19 laboratories, mainly spread throughout Europe but with one of Japan and another one of Australia . Five brands of thermocyclers were involved and cover the main devices in use. Even if in-depth statistical analysis of the results is still ongoing, there is no doubt that one can conclude the trial gives evidence that the designed protocol is fit for purpose with respect to the determination of a cut-off value.

Developing PCR on single particles is another idea worked out at CRA-W with the aim to take advantage of two techniques : on one hand, NIR-miscropy or imaging to make the distinction between authorized and unauthorized particles and then PCR on the other hand to define from which animal(s) the particle originates. A quick, easy and efficient protocol for directly performing the PCR on the particle was selected through comparisons of several protocols. It delivers enough material for five different PCR-tests. One point that was clearly evidenced is that a particle is not necessarily monospecies, it can be contaminated by tiny quantities of material of another species. Therefore purification steps of a particle are required to get rid of external contamination by material of other species so as to apply the test as much as possible on material of only one species. This is a much more tricky aspect still under development within the SA FE ED-PAP project.

Acknowledgements

G. Berben ¹
O. Fumière ¹
A. Marien ¹
V. Planchon ²
R. Oger ²

1 CRA-W, Department Quality of Agricultural products, Belgium