Outline on EVALLER™
The updated EVALLER™
Outline of DFLAP
Outline on EVALLER™
EVALLER™, which accomplishes computational risk assessment of a protein based on its amino acid sequence, is built on our recently published DFLAP (Detection based on Filtered Length-adjusted Allergen Peptides) algorithm (see below for more information on DFLAP). The user enters the amino acid sequence of interest, and when the DFLAP machine has processed the submitted amino acid sequence, EVALLER™ returns an assessment of its potential allergenicity/cross-reactivity, appearing as a comprehensive graphical and textual output. Each such assessment is accompanied by an assignment-specific uncertainty probability. Moreover, EVALLER™ supplies external links to detailed information on allergens most similar to the query protein, regardless of whether it is assigned as a potential allergen or a presumptive non-allergen. See also the link “How EVALLER™ works”.
The updated EVALLER™
Since publication of the article on EVALLER™ (Martinez Barrio, A. et al., 2007, Nucleic Acids Res, 35 (Web Server issue): W694-W700), a few improving amendments have been made on the server. Firstly, and most importantly, the current allergen data set is compiled from the AllergenOnline repository, January 2007 release, on allergen proteins (http://www.allergenonline.com/). A written permission to use AllergenOnline for the purpose of DFLAP updating was kindly provided by Prof. Richard Goodman, University of Nebraska, Food Allergy Research & Resource Program, Lincoln, USA. This enabled us to train and test the DFLAP algorithm with an appreciably larger and more extensively scrutinised repository, relative to our former catalogue. Secondly, the filter data set, employed to create a set of FLAPs (Filtered Length-adjusted Allergen Peptides), now encompasses solely the human proteome (Downloaded from Swall, 2007-05-08), which also was accessed much later than that referred to in our above-mentioned report. Altogether, these changes confer enhanced performance to EVALLER 2.0, thereby reducing likeliness for spurious results. Thirdly, the layout is markedly changed and accordingly harmonised with the National Food Administration’s overall graphical and textual profile.
Outline of DFLAP
The Detection based on Filtered Length-adjusted Allergen Peptides (DFLAP) algorithm is founded on a novel principle, which involves three main features:
- Firstly, a peptide-selection procedure from allergens is performed by comparing allergens with human proteins in a process that resembles filtration. The resulting peptide set is designated “Filtered Length-Adjusted Allergen Peptides” (FLAPs).
- Secondly, specific training of a support vector machine (SVM) - a supervised classification algorithm - by comparing similarity of known allergens and presumed non-allergens to the specially compiled peptide data set (FLAPs). The educated system is named Detection based on Filtered Length-adjusted Allergen Peptides (DFLAP).
- Lastly, a query protein’s unknown allergenic potential may be predicted based on its amino acid similarity to the FLAPs using DFLAP.
Validation test procedures have demonstrated high selectivity of the DFLAP system within protein families known to hold both allergens and non-allergens. Moreover, a low proportion of the SwissProt repository was assigned as allergens, indicating low false-alarm rate. For additional information on DFLAP and EVALLER™, please consult the link “Documentation”.