Version History History of the AllergenOnline Database Releases

The table below shows the total count of sequences in each release of the allergen online database. Also shown are the number of allergen groups (A collection of protein types within a taxonomic group) and number of unique species represented in each version.

Version Release Date Sequences Groups Species
5 Jan 2005 1189 608 208
6 Jan 2006 1537 672 255
7 Jan 2007 1251 461 221
8 Jan 2008 1313 483 229
9 Jan 2009 1386 502 236
10 Jan 2010 1471 529 254
11 Feb 2011 1491 553 265
12 Feb 7, 2012 1603 603 273
13 Feb 12, 2013 1630 612 275
14 Jan 20, 2014 1706 645 290
15 Jan 12, 2015 1897 744 335
16 Jan 27, 2016 1956 778 345
17 Jan 16, 2017 2035 808 354
18 Jan 18, 2018 2101 833 376
18A Feb 1, 2018 2093 832 370
18B Mar 23, 2018 2089 831 369
19 Feb 10, 2019 2129 853 384
20 Feb 10, 2020 2171 873 392
21 Feb 14, 2021 2233 913 430

Version 21: February 14, 2021

Note, version 21 was released on 14 February 2021. We will update the database during 2021 to align with issues for overly conservative CODEX 2003 guidelines for % identity matches. These should indicate different levels for various protein types.

New Feature: February 7, 2019

Added Allergenicity* to explain differences in "proof" of allergenicity as either "IgE binding but no biological test" or "IgE binding plus basophil+ or SPT+" results within the "Browse the Database" table.

New Feature: May 10, 2018

We have added information for each entry that states whether evidence was found indicating IgE binding as well as biological activity of basophil activation or skin prick test positivity for individual entries, or if the protein was only shown to have in vitro IgE binding. That information is available on the Browse page and it is intended to aid in a more accurate evaluation of protein risk, since there are tremendous differences in risk to different "allergens". It is important to note that proteins that do not have proof of biological activity may indeed be important allergens, but they have not been fully tested. Likewise, proteins with positive bioactivity may have been tested at very high doses. Review of references associated with the allergens are necessary to understand risk. Dose is also an important factor to consider.

Version 18B: March 23, 2018

NOTE 2: Removal of green-fluorescent protein from Scleronephthya gracillima on 23 March 2018.

In March, 2018 we were asked an important question regarding the validity of the allergenicity of the green-fluorescent protein from Scleonephthya gracillima an octocoral coral. There are four proteins of greater than 90% identity and only one paper, from Kato et al., 2017, Luminescence 32(6):1009-1016. Our panel had given it mixed reviews in accepting that as a putative allergen. Upon re-review 20 March, 2018, the full panel agreed the data from Kato et al., did not demonstrate specificity of IgE binding. Their patient population was unspecificied lobster fishermen from Japan with dermatitis or asthma. Pooled sera was used to demonstrate IgE binding. There was markedly more binding to two other uncharacterized proteins, altough the major IgE binding protein also had binding from a pool of "non-allergic" subjects. The protein sequence was determined by a peptide from LC-MSMS, then cDNA clones were constructed. The allergic subjects sera was not used to demonstrate IgE binding to pure protein. We conclude that the IgE binding was not specific or that the binding was sufficiently weak and the protein should not be included even as a "putative" allergen, our lowest category of allergen. A more detailed report can be obtained by email request to

Letter GFP removal

Version 18A: January 18, 2018

Note, version 18 update was released on 18 January 2018, but was missing some information of WHO/IUIS allergen numbers and also used a few accession numbers that were not optimum. A full review of the database was performed from 19 January until 31 January with corrections made as version 18A which was posted on 1 February 2018.

Celiac Database

A Beta version of the Celiac database was installed in August, 2017. The updates included new 9 amino acid peptide entries from the European Food Safety Authority panel as they should be bound by HLA DQ2 (DQ2.5, DQ2.2, Cis or trans) or DQ8. Beta version 2 was updated in September, 2017 with removal of 8 amino acid matches, bringing the total to 2013 peptides including native and specifically deamidated peptides. BLASTP testing with the 9 AA peptides demonstrated that a few match proteins outside of the grass family and the word "CAUTION" was added for those entries. Beta version 3 was posted on 5 October 2017 and includes four new representative proteins. The FASTA search feature has been validated again and the final Celiac Disease database version 2 release was 18 January 2018. RE Goodman

Version 1 of the Celiac Database was available for public use on 14 February, 2012.

Version 2: January 18, 2018 Peptides, Proteins and References

This database has been updated for prediction of risks of stimulating celiac disease (CD) for CD subjects. The update includes removal of peptides less than 9 amino acids (AA) long as they are too small to interact with MHC-II and T-cells and are therefore not predictive. This update includes addition of all peptides included in the July 2017 EFSA Appendix A Allergenicity update (EFSA Journal 15(6):4862) as well as most of the peptides from our 2012 Celiac Database (1016 peptides) from 2012 previously on this website. The peptides include native and predicted deamidated peptides of wheat, barley, rye and oats that are associated with CD. NOTE the EFSA list of 9 AA peptides are considered strong candidates for binding MHC Class II HLA DQ2 or DQ8, but that includes 30% of the general population. Since CD effects approximately 1.3% of the global population, the specificity of T-cell receptors for longer or substituted AA positions will alter specificity. In addition other genes or environmental factors (gut bacteria, virual infections) are also important in stimulating development of CD. This database is intended as a risk assessment tool for developers of genetically engineered (modified) crops and novel food ingredients, as well as for researchers. Therefore exact identity matches to a 9 AA peptide needs to be interpreted as it will likely over-estimate peptides that pose a risk of eliciting CD. We include 1013 peptides from 9 AA to 53 AA that are published as binding HLA DQ 2 or DQ8, or of stimulating CD T-cell proliferation or toxicity using human samples or subjects. We have therefore included representative proteins and a FASTA algorithm for further analysis. Developers finding matches to the peptides and proteins as discussed in detail here, are encouraged to label their products as containing "gluten" unless they conduct further research to demonstrate their protein will not elicit CD. The references are updated. Search tools remain the same. The peptide entries were reviewed by an academic committee. Proteins will be updated following additional testing. The guidance and suggested criteria for significance is currently being tested and reviewed and will be modified by January 2018. Special thanks to Plaimein Amnuaycheewa, Frits Koning and Barbara Bohle for their contributions. RE Goodman, 18 January, 2018.