3.2—Today’s food allergy criteria are more accurate

Improper testing criteria can make any protein appear to resemble an allergen

See Genetic Roulette’s False Claims at Bottom of Page

Analysis of Peer-Reviewed Research:

Genetic Roulette claims that regulators have approved crops that contain proteins that are potentially allergenic. The problem with the book’s claims is that they are based on unreasonable and now discredited rules of thumb that have been used in the past to evaluate allergenicity.  The chances that a protein will be an allergen are vanishingly small if it does not closely resemble known allergens.  Thus, detailed and comprehensive comparison of a protein to all known allergens provides much insight into its potential to be an allergen.  Obviously, if we change the rules so that less similar proteins are also scored as allergens, many more proteins will be red-flagged as potential allergens.  This does not mean that they really are allergens, but simply that we have picked very unreliable rules for red-flagging the protein.  Smith doesn’t tell the reader about the latest research on allergenicity that shows his examples shouldn’t be considered potential allergens.

1.  Tests can assure that a GM protein is no more likely to be an allergen than any other protein in our diets. Most food allergens are proteins (Lehrer and Bannon 2005; Mills and others 2004).  A single food can contain thousands of different proteins so that humans around the globe probably consume millions of different proteins.  Very, very few of these proteins will cause food allergy.  In fact, 10 major foods cause 90 percent to 95 percent of all food allergies, and there are perhaps 200 or more foods that rarely cause allergies (Hefle and others 1996).  The structure and exact chemical configuration of almost all of these allergens has been determined and is available in databases (see for example: www.allergenonline.com/).  Scientists can compare a protein to the entries in these databases and with computer software assistance determine if the protein resembles a known allergen.  The objective of this safety assessment process is to ensure that no proteins that resemble known allergens are incorporated into foods.  These tests cannot predict an unknown or previously undescribed allergen, but because of the analysis and testing that are done, we can say that a GM protein is no more likely to become a new allergen than is any other protein in our diets (Goodman and others 2008, Ivanciuc and others 2008).

2.  How similar a protein is to an allergen depends directly on how you define similarity. If you set the criteria by which one protein is compared to all allergens too restrictively, only identical proteins would score a match.  If you set the criteria for a match too loosely, many proteins that are not allergens would be scored as possible allergens.  Much research and analysis has gone into finding criteria that allow us to conclude with confidence that a protein does or does not closely resemble an allergen (Goodman and others 2008, Ivanciuc and others 2008).

3.  FAO and WHO wanted to ensure that all potential allergens would be detected so they lowered the criteria for an allergen match to a point that many non-allergens were predicted to be allergens. Smith does not bother to tell the reader that there has been a long debate among scientists about how stringently (how narrow) the criteria for scoring a positive hit should be set (FAO/WHO 2001, Goodman and others  2008, Ivanciuc and others 2008).  A bit of information about food allergens is necessary to understand what happened in this case.

Food allergy is mediated with a class of antibodies called IgE antibodies.  These antibodies have specific regions that bind to specific target sites on the allergen – think of these as two pieces in a puzzle that fit together.  No allergic reaction takes places unless an allergen molecule binds to its IgE antibody at its specific binding site.  These binding sites for antibodies on allergens are called epitopes and the distinctive pattern of protein building blocks (called amino acids; the units from  which all proteins are composed) at the epitope – called the amino acid sequence – has been determined for many allergens.

Think of the epitopes as the fingerprints of an allergen.  No two are identical and they must match their own IgE antibodies to function.  Using knowledge of these sequences, and a computer search program, one can determine if a target protein contains any known epitope—it’s just like looking for a fingerprint match.  Obviously if it contained an epitope found in an allergen, it would signal the need for further study.  The Devil, as they say, is in the details of deciding just how similar a sequence needs to be to an allergen and send up a red flag.

For many years scientists used the rule of thumb that if they found eight amino acid building blocks in a row that were identical to an eight building block epitope existing in any known allergen  they would flag the protein for further analysis.  It is critical to understand that a match does not mean that the protein is an allergen—it means possible similarity that should be studied further. Smith badly misrepresents the meaning of these kinds of analysis by claiming a match shows that a protein is an allergen.  An expert panel convened by FAO/WHO recommended in 2001 (FAO/WHO 2001), without having done any analysis of the consequences or implications, that the size of the search window should be lowered to six amino acids.  Their motives are unclear to us but appear to have been based on a politically motivated desire to add an extra-measure of precaution.

4.  The six amino acid window of building block identity recommended in 2001 by FAO/WHO has been shown to be an invalid method. Smith persists in calling proteins that have six amino acid epitope matches allergens.  To begin with, these matches don’t prove a protein is an allergen; they are one measure of similarity that is considered along with other information (Goodman and others  2008, Ladics and Selgrade 2008).  More importantly, it is now clear that an unacceptable number of false protein classifications are found when a six amino acid window is applied.  One can calculate that patches of six identical amino acids in a row will occur far more often in a protein molecule that is 500-1000 amino acids in length than will patches of identical eight amino acids in a row (remember that a run of four aces in a poker hand is less common that three aces).  So frequently do six building block matches appear in proteins that 85 percent of all corn proteins studies are red-flagged as potential allergens using a six building block in a row identity as a rule of thumb.  The six amino acid window has been demonstrated to be an inappropriate criteria for identification of potential allergens, and it has been suggested by reputable allergy specialists that the FAO/WHO 6-mer recommendation should be disregarded (Ladics and others 2006, Silvanovich and others 2006; Goodman and other  2008, Ivanciuc and others 2008).  Similar questions have been raised about the use of other criteria recommended by the FAO/WHO panel (Cressman and Ladics 2008; Ladics and Selgrade 2008).

The take-home message here is that even experts can be wrong when they become too precautionary (and political?) and don’t pay attention to sound science.  Smith obviously hasn’t abandoned this obsolete concept either because he hasn’t read the literature or because it allows him to claim that regulatory agencies knowingly approve potential allergens.  Smith gives other examples in which the analogy to a known allergen was not sufficiently close to be a reason for concern.

5.  Protein digestibility is one of many factors considered when allergenicity is tested—many indigestible proteins are not allergens. Smith claims that various proteins in GM crops that have failed the digestive enzyme test—that is to say they do not digest or do not digest rapidly enough.  He also claims that the test itself is flawed in that it does not reflect what happens in the gut.  One problem with the digestibility test has been lack of standardization between laboratories until recently, when a large group of laboratories joined together to test a protocol that everyone could use (Thomas and others 2004).  The proteins in currently planted GM crops are considered digestible by this standardized method.  It is not claimed that this method works exactly like the human digestive tract—it probably doesn’t.  Regulators use the test to provide an additional layer of assurance that a protein is safe since rapidly digestible proteins are less likely to be allergens.  The predictive value of the protein digestibility test itself has recently been called into question (Goodman and others 2008). Genetic Roulette claims that proper attention has not been paid to another test, susceptibility to heat.  Regulatory authorities do not require heat stability testing because it would not be predictive of safety in any known way. No single test is used to evaluate allergy potential – a weight of all the evidence is used.

6.  Some of the proteins that are claimed in Genetic Roulette to be allergens are used in large quantities around the world and no allergies have been reported. Proteins can be divided in a number of families whose members closely resemble one another (Mills and others  2004; Ivanciuc and others 2008).  We now know that food allergens are almost always members of one of three large families of proteins.  If a protein is not a member of one of these three families, it is highly unlikely to ever become an allergen.  None of the proteins used in commercial GM crops — without exception — are members of the families of proteins that encompass the great majority of the plant food allergens.  Regulators are well aware that evaluating allergenicity is important and that a protein must resemble other known allergens very closely in several ways before it poses a threat.  Regulators do not approve crops when there is any residual concern about allergenicity.  If anything, history shows that they are willing to err on the side of safety if there is the slightest doubt about a protein’s potential allergenicity.

(link to more about allergenicity: www.foodallergy.org/, www.nlm.nih.gov/medlineplus/foodallergy.html, www.faiusa.org/?&CFID=11502380&CFTOKEN=32132995)

See also

3.5 The Pesticide in Starlink has a very low likelihood of being an allergen


Cressman RFand Ladics G (2008). Further Evaluation of the Utility of “Sliding Window” FASTA in Predicting Cross-Reactivity with Allergenic Proteins, Regulatory Toxicology and Pharmacology (2008), doi: 10.1016/j.yrtph.2008.11.006

FAO/WHO (2001). Evaluation of allergenicity of genetically modified foods. Report of a joint FAO/WHO expert consultation on allergenicity of foods derived from biotechnology. (Food and Agriculture Organization of the United Nations (FAO), Rome, 2001.  ftp.fao.org/es/esn/food/allergygm.pdf

Goodman RES, Vieths S, Sampson HA, Hill D, Ebisawa M, Taylor, SL and van Ree R (2008). Allergenicity assessment of genetically modified crops: what makes sense? Nat. Biotech. 26: 73-81.

Hefle SL, Nordlee JA and Taylor SL (1996). Allergenic Foods

Critical Reviews in Food Science and Nutrition. 36(S):S69-S89

Ivanciuc O, Garcia T, Torres M, Schein CH and Braun W (2008). Characteristic motifs for families of allergenic proteins. Molecular Immunology.
doi:10.1016/j.molimm.2008.07.034 <dx.doi.org/10.1016/j.molimm.2008.07.034>

Ladics GS and Selgrade MK (2008). Identifying food proteins with allergenic potential: Evolution of approaches.  Regul. Toxicol. Pharmacol. doi:10.1016/j.yrtph.2008.10.010

Ladics GS, Bardina L, Cressman RF, Mattsson JLsnd  Sampson HA (2006). Lack of cross-reactivity between the Bacillus thuringiensis derived protein Cry1F in maize grain and dust mite Der p 7 protein with human sera positive for Der p 7- IgE. Regul. Toxicol. Pharmacol. 44: 136–143.

Lehrer SB andBannon GA  (2005). Risks of allergic reactions to biotech proteins in foods: perception and reality. Allergy, 60: 559–564

Mills ENC, JA Jenkins JA, Alcocer MJC, Shewry PR (2004). Structural, biological, and evolutionary relationships of plant food allergens sensitizing via the gastrointestinal tract. Critical Reviews in Food Science and Nutrition, 44:379–407

Silvanovich A, Nemeth MA, Song P, Herman R, Tagliani L and Gary A. Bannon GA (2006).  The value of short amino acid sequence matches for prediction of protein allergenicity. Toxicological Sciences 90:252–258

Thomas K, Aalbers M, Bannon GA, Bartels M, Dearman RJ, Esdaile DJ, Fu TJ, Glatt CM, Hadfield N, Hatzos C, Hefle SL, Heylings JR, Goodman R E, Henry B, Herouet C, Holsapple M, Ladics GS, Landry TD, MacIntosh SC, Rice E A, Privalle LS, Steiner HY, Teshima R, Van Ree, R, Woolhiser M, and Zawodny J (2004). A multi-laboratory evaluation of a common in vitro pepsin digestion assay protocol used in assessing the safety of novel proteins. Regul Toxicol Pharmacol 39:87-98.

Genetic Roulette Falsely Claims:
GM proteins in soy, corn, and papaya may be allergens.

1. Tests cannot guarantee that a GM protein will not cause an allergy.

2. The World Health Organization (WHO) and the U.N. Food and Agriculture Organization (FAO) offer criteria that help minimize the likelihood that allergenic GM crops are approved.

3. GM soybeans, corn, and papaya fail those criteria.

4. The GM proteins from these foods are too similar to known allergens.

5. This evidence was ignored by regulators, who approved the crops.

The WHO and FAO recommended criteria that make many proteins appear similar to allergens.