Introduction

Test procedures for allergenic potential
Development of bioinformatics testing methods
EVALLER: built on a specially compiled set of peptides and statistical learning
Selected literature

Test procedures for allergenic potential

Various immunochemical, biochemical and immunological methods, dedicated to identification of proteins with a potential to cause IgE-mediated allergic reactions, have arisen and evolved over time. They include IgE immunosorbent assays using patient blood sera, human skin prick tests, double-blind placebo-controlled food challenge (DBPCFC), in vitro basophil histamine release as well as various animal models, the first three assay methods also being used in clinical diagnostics.

Bioinformatics test procedures for allergenicity have attracted mounting interest among risk assessors in toxicology and allergology. Largely, this direction has emerged in response to a need for a relatively robust, easy and swift approach to screening for potential protein allergens, due to concern over the possibility of unpremeditated introduction of allergen-encoding xenogenes into genetically modified (GM) food crops. Nonetheless, this sort of expedient interrogation for allergenic potential should also apply to other areas, typified by Novel Foods (as defined by the (EC) 258/97 Regulation) and biopharmaceuticals.
 

Development of bioinformatics testing methods

Allergy involves a range of diverse mechanisms, which has urged recommendations to using various different methods, including those referred to above, to safely conclude on potential protein allergenicity. Bioinformatics-type inspection, though, represents a key procedure for allergenicity testing in the Codex Alimentarius guideline on safety assessment of genetically modified foods as well as in a cognate instruction compiled by the European Food Safety Authority (EFSA).

In essence, the Codex protocol prescribes a two-part in silico interrogation practice wherein potential hazard is signalled by either a short (but not specified in length) match of consecutive amino acids or an identity of more than 35 %, over an 80-amino acid window, in both cases to a documented protein allergen. The latter part of this combination procedure has emerged as the most robust among them.

The complex relationships between amino acid sequence similarity to known allergens and allergenicity/cross-reactivity potential have, however, prompted further development within this particular field of risk assessment. The implementation of several mutually dissimilar comparison approaches, in conjunction with statistical learning algorithms, has conferred increasing overall performance to computational allergenicity assessment.
 

EVALLER: built on a specially compiled set of peptides and statistical learning

We have chosen alignment-based feature-extraction in conjunction with supervised learning as an inroad to bioinformatics-type assessment of protein allergenicity. By bringing a special kind of segment-reduction procedure into action on allergens, the alignment/supervised learning model - ultimately appearing as DFLAP (Detection Based on Filtered Length-adjusted Allergen Peptides) - has been considerably enhanced. Moreover, the DFLAP SVM (Support Vector Machine) algorithm, being the core of EVALLER™, is furnished with new training and validation data since publication of the web server in 2007. A special report on EVALLER™ 2.0, appearing in a 2007 issue of EMBnet.news journal, is recommended to the interested reader:

Bongcam-Rudloff, E., Edsgärd, D., Martinez Barrio, A., Soeria-Atmadja, D., Gustafsson, M.G. and Hammerling, U. (2007) A guide to EVALLER™ (2.0) web server: A new tool for in silico testing of protein allergenicity.  EMBnet.news 13 (4): 32-37.

Selected literature

1. Codex Alimentarius Commission (2003) (ALINORM 03/34A) Guideline for the conduct of food safety assessment of foods derived from recombinant DNA plants. Annex on the assessment for possible allergenicity, Rome, Italy.
2. EFSA (2004) Guidance document of the Scientific Panel on Genetically Modified Organisms for the risk assessment of genetically modified plants and derived food and feed. The EFSA Journal 99, 1-94.
3. van Ree, R., Vieths, S. and Poulsen, L.K. (2006) Allergen-specific IgE testing in the diagnosis of food allergy and the event of a positive match in the bioinformatics search. Review of the development of methodology for evaluating the human allergenic potential of novel proteins. Mol Nutr Food Res 50 (7), 645-654.
4. Soeria-Atmadja, D., Lundell, T., Gustafsson, M.G. and Hammerling, U. (2006) Computational detection of allergenic proteins attains a new level of accuracy with in silico variable-length peptide extraction and machine learning. Nucleic Acids Res 34 (13), 3779-3793. 
5. Kong, W., Tan, T.S., Tham, L. and Choo, K.W. (2007) Improved prediction of allergenicity by combination of multiple sequence motifs. In Silico Biol 7, 0006. 
6. Schein, C.H., Ivanciuc, O. and Braun, W. (2007) Bioinformatics approaches to classifying allergens and predicting cross-reactivity. Immunol Allergy Clin North Am 27 (1), 1-27.