Practical and predictive bioinformatics methods for the identification of potentially cross‐reactive protein matches

Abstract

A bioinformatics comparison of proteins introduced into food crops through genetic engineering provides a mechanism to identify those proteins that may present an increased risk of allergic reactions for individuals with existing allergies. The goal is to identify proteins that are known to be allergens or are so similar to an allergen that they may induce allergic cross‐reactions. Three comparative approaches have traditionally been used, or considered for safety evaluations. One identifies any short (6–8) amino acid segment of the protein that exactly matches a known allergen sequence. The second is an overall primary sequence comparison using Basic Local Alignment Search Tool (BLAST) or FASTA to find matches of greater than 35% identity over 80 amino acids. The third is based on 3‐D prediction programs to identify 3‐D similarities that might predict potential cross‐reactivity. The utility of each of these approaches was debated in the bioinformatics workshop. The consensus agreement from the expert workshop participants was that the short‐segment match (e. g., 6–8 amino acids) provides an unacceptably high rate of false positive matches and an uncertain rate of true positive matches, and was not particularly useful for an allergenicity evaluation performed in the context of comprehensive safety evaluation. There was no consensus regarding the most appropriate bioinformatics method, an acceptable scoring criteria for triggering closer examination subsequent to a positive match, or an acceptable scoring mechanism for ranking the utility of the various 3‐D approaches that were discussed during the workshop. However, the general consensus was that the most practical approach at this time is to evaluate primary sequence identities to known allergens using either FASTA or BLAST. While there was good agreement that identities of greater than 35% over 80 or more amino acids (recommended by Codex in 2003) is quite conservative, the conclusion was that additional data or studies would be needed to justify changing this criterion as there is some evidence that some individuals sensitized to proteins in evolutionarily conserved protein families may experience cross‐reactions to proteins sharing approximately 40% identity.

The Occurrence of Sequences Identical with Epitopes from the Allergen Pen a 1.0102 Among Food and Non-Food Proteins

“…The remaining parameters were set at default values. Only the sequences whose identity 100% matched the query sequences or the length of continuous fragment exceeding 8 amino acid residues were taken into account in line with previous recommendations [Goodman, 2006;Minkiewicz et al, 2011]. Vertebrate protein sequences annotated as "inferred from homology" or "predicted" were examined using the MisPred program [Nagy et al, 2008].…”

“…Pentapeptides are the shortest fragments recognised by the immune system [Kanduc, 2008]. The existence of a common fragment containing at least 6-8 amino acid residues is a recommended bioinformatics criterion defi ning protein as an allergen cross-reacting with previously-known allergenic proteins [Goodman, 2006]. Protein fragments used usually for epitope mapping have at least 10, and usually 15, amino acid residues.…”

Allergenicity Assessment of a Genetically Modified Protein-Recombinant Human Lactoferrin

“…The amino acid sequence comparison for rhLF was conducted with three databases; FARRP (http://www.allergenonline. com), SDAP (http://fermi.utmb.edu/SDAP/sdap_src.html) and ADFS (http://allergen.nihs.go.jp/ADFS/), which are widely used in the world at present [40][41][42]. The primary methods of evaluation were searches for "80 amino acid alignments with greater than 35% identity by FASTA" and "8 amino acid exact matches".…”

Evaluation of In Silico Prediction Possibility of Epitope Sequences Using Experimental Data Concerning Allergenic Food Proteins Summarized in BIOPEP Database.

“…: +4889 523 37 15 E-mail: jerzy.dziuba@uwm.edu.pl (Prof. J. Dziuba) Health Organization (WHO), such as the presence of protein sequence fragments containing a minimum of 6-8 amino acid residues which are identical to the fragments of known allergens or fragments containing a minimum of 80 amino acid residues showing at least 35% similarity with the known allergen sequence [Goodman, 2006]. The search for new solutions is still necessary to further advancement of the existing bioinformatics methods and tools [Gowthaman & Agrewala, 2009].…”

“…The algorithm applied to develop the EVALLER program was compared with other algorithms investigating the allergenic character of the examined proteins [Soeria-Atmadja et al, 2006]. In reference to the offi cial bioinformatics criteria recommended by the WHO [Goodman, 2006], the discussed program produces fewer false-positive results, i.e. cases in which a non-allergenic protein is found to be an allergen [Soeria-Atmadja et al, 2006].…”