you are in: Home > The Conference > press info


For Press Correspondents
Thank you for attending The GMSAFOOD Project’s March 8th Press Conference. Photos are available on our website, with more to be added daily. A post-conference newsletter in in the works, and we are also editing footage, audio recording, and power point presentations of conference talks, to be made available with the speaker's permission. If there is anything you require, or any way in which I may assist you, please do not hesitate to ask.

Below, I would like to provide some additional information regarding our findings in an effort to further illuminate the import of what we have presented.

Firstly, to better explain the significance of our proposal for a machine-learning approach to post-market monitoring. The unique value of our approach is its ability to potentially identify a unifying biomarker that could offer new insights into the long-term effects of novel food consumption across species. (A biomarker is any physiological indicator of a particular biological state or condition – for example, high blood pressure is a biomarker useful in assessing risk of heart disease.) Our proposed approach is neither traditional meta-analysis nor epidemiological post-market monitoring.

Traditional meta-analysis is limited to combining the results of highly similar studies. If studies are not sufficiently relevant, no meaningful analysis can be conducted. This new approach is not similarly limited. Data from extremely different experiments can be meaningfully related, revealing previously hidden common underlying factors. Additionally, results of traditional meta-analysis can be completely derailed by a poorly designed systematic review. Because our approach is hunting for a robust signifier among a multitude of disparate studies, we hypothesize that misleading results are highly unlikely to emerge from within the “noise” of unrelated or insignificant data – even if a study that is later discredited by peer review is initially included in our analysis.

As in traditional meta-analysis, this approach will help to overcome “presentation bias” – instead of a single study being afforded undue weight or influence due to bias, publication, or dissemination. Studies with less than sensational results can be overlooked in exchange for a study that confirms one’s beliefs, potentially leading to public misperception. This approach would not contribute to, and is itself free from such influence.

This is also distinct from traditional epidemiological post-market monitoring, which is limited to exploring previously identified health effects. Unlike epidemiological PMM, which must be hypothesis-driven, the value of our approach is its ability to uncover unforeseen health factors. To date, no health risks have been definitively linked to any novel food. The machine-learning approach addresses many of the challenges and limitations of epidemiological post-market monitoring, such as the difficulties in monitoring use within the general public, and assessing the association between the introduction of a substance into the food supply and health effects. This proposed approach would complement, rather than replace, current pre- and post- market surveillance techniques.

Also worthy of note are the contradictory findings of 2005’s Prescott et al. study and the GMSAFOOD consortium’s more recent study. The Prescott study observed an allergic response in mice to peas modified to produce an alpha-amylase inhibitor protein (AAI), but not in the bean which naturally produces AAI. This bean, a variety called Tendergreen, was the source of the genetic material for producing AAI in the GM pea. Conversely, the GMSAFOOD study observed an allergic response to not only the GM pea, but also the Tendergreen bean, and even to the non-GM pea (which contains no AAI). While more experiments must be conducted to explain the conflicting results, the new GMSAFOOD studies suggest caution for researchers utilizing mice as models for investigating human allergenicity to GMOs or other novel foods.


Thank you for your attendance and interest in our project!


Stephen Doyle

Manager of Public Relations


Tel.: +43 1 40160 63009