http://primalseeds.nologic.org/dream.htm

Because no gene ever functions in isolation, there will almost always be unexpected and unintended "side-effects" from the gene or genes transferred into an organism.

One major concern over transgenic foods is their potential to be toxic or allergenic, which has become a concrete issue since a transgenic soybean containing a brazil-nut gene was found to be allergenic. Recent studies suggest that allergenicity in plants is connected to proteins involved in defence against pests and diseases. Thus, transgenic plants engineered for resistance to diseases and pests may have a higher allergenic potential than the unmodified plants.

New proteins from bacterial, such as the Bt toxin currently engineered into many transgenic crops, cannot be tested for allergenicity because allergic reactions depend on prior exposures. That means post-market monitoring and clear segregation and labelling of transgenic products are essential for proper consumer protection. Most identified allergens are water-soluble and acid- resistant. Some, such as those derived from soya, peanut and milk, are very heat-stable, and are not degraded during cooking, whereas fruit-derived allergenic proteins are heat-labile.

A transgenic yeast engineered for increased rate of fermentation with multiple copies of one of its own genes, which resulted in the accumulation of the metabolite, methylglyoxal, at toxic, mutagenic levels. This case should serve as a warning against applying the "familiarity principle" in risk assessment. We simply do not have sufficient understanding of the principles of physiological regulation to enable us to categorize, a priori, those genetic modifications that will pose a risk and those that do not.