Introduction
EVALLER TM is a web server that allows for bioinformatics assessment of proteins in the context of allergenicity (type-I hypersensitivity). The server is built on principles and directions described in published articles 1-3 and particularly on the DFLAP (Detection based on Filtered Length-adjusted Allergen Peptides) core algorithm 3.
DFLAP is constructed on an algorithm that compares query proteins with a specially compiled peptide data set, using judgements based on statistical learning.
The peptide data set was derived by filtering allergen amino acid sequences in silico through a large set of proteins with low probability of being allergenic (largely the human proteome). 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 a low false-alarm rate.
EVALLER TM accepts a single amino acid sequence in FASTA format, which should either be pasted into the acceptor window or uploaded as a file. The output includes a two-category assignment (presumptive allergen or presumptive non-allergen) as well as a scoring value and a ratio number. These numbers, and particularly the latter one, are provided to inform on the degree of assessment certainty for a DFLAP score value. The posterior probability ratio, directly derived from the DFLAP score value, expresses a confidence-related figure, i.e. an index of categorisation firmness. Phrased in greater detail, the ratio involves computation of the probability for being either allergen or non-allergen based on output of a set (1175 entities) of training amino acid sequences that were fed to EVALLER TM. Due to differences in distribution profiles of these two categories the maximal ratio for non-allergens is 150, whereas those of allergens rapidly escalate to very higher numbers. Moreover, the algorithm will inappropriately assign a gradually lower ratio to non-allergens with unusually low score values. Therefore, the EVALLER TM output on ratios will rise to limits, beyond which 150 (max) and >1000, respectively, are displayed.
To further support risk assessment in allergology EVALLER TM also provides a more comprehensive synopsis of its judgement on a query proteins potential allergenicity. Each such textual/graphical output includes information on the best match(es) against allergen-related motifs, including their respective location in accordingly derived query proteins. This feature should be most relevant for allergens, but is also allowed to operate on proteins falling in the category that signifies low probability of allergenicity, as an extended support to the viewer in ambiguous cases. The latter feature involves colour-coded graphic representation (red/green) of amino acid sequences matches to known allergens. The interface is dynamic and enables examination of pertinent details in great detail through zoomed views.
Distributed and published results, directly or indirectly derived from EVALLER TM usage, should accompany reference to Soeria-Atmadja, G. et al., 2006 3.
Literature:
1 Bjorklund, A.K., Soeria-Atmadja, D., Zorzet, A., Hammerling, U. and Gustafsson, M.G. (2005) Supervised identification of allergen-representative peptides for in silico detection of potentially allergenic proteins. Bioinformatics 21: 39-50.
2 Soeria-Atmadja, D., Wallman, M., Bjorklund, A.K., Hammerling, U. and Gustafsson, M.G. (2005) External cross-validation for unbiased evaluation of protein family detectors: Application to allergens. Proteins 61: 918-925.
3 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: 3779-3793.