Prédiction in silico d'épitopes d'allergènes alimentaires et validation expérimentale sur les allergènes du blé – PREDEXPITOPE

"Food allergies represent a major and growing health-issue in western societes. Their are induced by numerous food products as egg, milk, treenuts and vegetables and evolve with feeding customs. According to the age, they elicite several symptoms like atopic dermatitis, asthma and anaphylactic shock. They account for 0.2-1% of ermergency consultations and this has significantly increased for these past few years. The only treatment of food allergies relies on an eviction of the offending foods, which is often constricting, specially when it's based on food ingredients. Localization of antigenic determinants of allergens would help to better understand allergy mechanisms, identify risk factors an delaborate preventive strategies.
Unfortunatly, the current experimental techniques are expensive and heavy to establish, preventing thus an extended approach to study a large scale of allergens. In such a context, the in silico preselection of putative epitopes is desirable. This project aims at including structural " "bioinformatics approaches to identify immunogenic areas of allergens, predicting their digestibility according to their folding level and their digestive enzymes cutting sites, and also at studying cognate epitopes proximities between each others (B-cell adjacent epitopes and T-cell epitopes proximity). These step imply to develop an allergen epitope database from public databases, completed by literature. This database helps not only to assess exsiting prediction tools but to deduce statistical rules from several parameters using learning methods.
Wheat allergy is one of the few frequent of food allergies and is yet less studied. It implies gliadin, glutenin and proteins from the albumin/globulin fraction. Few epitopes have already been idenitfied on these molecules composing thus a validation set for the prediction system developped here. This validation will be about linear epitopes, on the one hand and conformational epitopes, on the other hand. Linear epitopes will be identfied by antibodies binding test to overlapping synthetic peptides. Predicted conformational epitopes will be produced as long peptides and " "assessed by their ability to interact with IgE. Site-directed mutagenesis approaches will also applied to recombinant proteins or to their fragments. Furthermore, the allergy progression and the epitope prevalence will be studied following a cohort of patients, all along the project, and following also the allergens and peptides translocation through the epithelial hence.
All experimental results will be included to the epitope database and the final prediction system will be published as a software."