IPD-IMGT/HLA releases and why allele names change

Q: Where can I see which names changed in a release?

Allelelist_history.txt and Deleted_alleles.txt in the IPD-IMGT/HLA repository record every change. GET /v1/releases/diff summarises the differences between any two releases.

Updated 2026-10-29 · allele examples checked against IPD-IMGT/HLA 3.66.0 on 2026-10-29

The IPD-IMGT/HLA Database is the reference list of named HLA alleles, and it publishes a new release every quarter. Each release adds names, and a smaller number of existing names change: they gain a field, gain or lose an expression suffix, are renamed into another family, or are deleted.

So an allele name is only fully meaningful together with the release it was checked against. A typing that was correct when it was reported can contain a name that is no longer current today.

How release numbers work

The IPD-IMGT/HLA README explains the three-part release number, for example 3.66.0:

Part	Example	Meaning
First	`3`	The nomenclature version. Version 3 is the colon-separated nomenclature introduced in April 2010
Second	`66`	The quarterly release, incremented every January, April, July and October
Third	`0`	The sequence version: 0 for the primary quarterly release, incremented only if a later patch changes a valid base in a sequence

The data are published in the ANHIG/IMGTHLA GitHub repository. At the time of writing the most recent tag is v3.66.0-alpha, dated 14 October 2026, and the hla.alleles.org changelog lists Release 3.66.0 on the same date. The README also describes a separate testing branch for upcoming file-format changes; its files are not for clinical or diagnostic use and contain no unreleased names.

How much changes each quarter

Between release 3.65.0 and 3.66.0, the name history records 772 names added, none removed outright and 24 renamed. The hla.alleles.org statistics for September 2026 count 46,146 named HLA alleles in total, 31,311 class I and 14,835 class II, plus 1,278 alleles of related non-HLA genes.

Most of each release is new alleles. The renames are few, but they are the part that breaks stored data.

The kinds of name change

A field is added

A name is only as long as it needs to be to separate an allele from its relatives. When a new allele is found that differs from an existing one only by a synonymous change, the existing allele gains a third field. In release 3.66.0, A*01:284 became A*01:284:01 when A*01:284:02 was named. The two-field name A*01:284 is still a valid prefix, but a system matching exact strings will no longer find it in the current allele list.

A suffix is added or removed

Expression suffixes record the evidence at the time. The suffix page says assignments may be revised as new data arrive. One example from the name history: A*02:01:14 became A*02:01:14Q in March 2011, and the Q was removed in August 2026, making it A*02:01:14 again in release 3.66.0.

A name moves to another family

Occasionally an allele is renamed into a different first-field group. DRB3*03:22, later DRB3*03:22:01, became DRB3*02:173:01 in release 3.46.0. In release 3.66.0, DQA1*01:209N, which existed for one release only, became DQA1*03:100N; the deletion note says the sequence was "shown to belong to DQA1*03 allele group".

A name is deleted

Some names are withdrawn. Deletion notes in the file include sequences "shown to contain errors and be identical to" an existing allele, and sequences shown to belong to another allele group. Deleted_alleles.txt records each deletion with its reason and, where there is one, the successor name. The file also records suffix changes, which is why A*02:01:14Q appears there.

The files that make tracking possible

The README lists the files that carry history, all in the repository root:

Allelelist.txt: every allele named in the current release.
Allelelist_history.txt: each allele's official name in every release, keyed by a permanent HLA ID such as HLA02908.
Deleted_alleles.txt: deleted names with reasons and successors, including suffix changes.
release_version.txt: the current release number.

The HLA ID is the key to everything. Names change; the ID behind them does not. To find out what a stored name means today, look up its ID in the history, then read the ID's name in the current release.

The data are licensed CC BY-ND with a citation requirement, so cite the database when you use it.

Keeping stored typings current

A practical policy for a LIMS, registry or research dataset:

Store the release with every typing. Record the release the lab reported against, or the date if the release is unknown.
Never overwrite the reported string. Keep it next to any normalised form.
Re-check on every release. After each quarterly release, compare your stored names with the new allele list and the history.
Separate "renamed" from "deleted" from "unknown". Renamed names have a current equivalent. Deleted names may have a successor. Unknown names were never assigned and are data errors.
Re-expand groups. G and P groups are recalculated each release, so a stored group name can cover different alleles now. See G groups and P groups.

For archives that predate April 2010, see converting pre-2010 names, where the name history matters even more.

How HLA-Verify handles releases

Every HLA-Verify response names the release it was computed from. The deployment is pinned to one release at a time, currently 3.66.0, and the release history it carries covers 111 releases from 1.05.0 to 3.66.0.

Three calls cover release tracking:

GET /v1/releases/diff?from=3.65.0&to=3.66.0 lists every name added, removed and renamed between two releases, optionally for one locus.
POST /v1/translate maps a list of names written in one release to their names in another, through the HLA ID.
GET /v1/allele/{name}/history returns every name an allele has had, release by release.

What changed at each move of the pin is published at /releases, and the release translator runs the same calls in the browser.

Frequently asked questions

How often is IPD-IMGT/HLA released?

Quarterly. The README says the release number's second field is incremented every January, April, July and October.

What does the 3 in 3.66.0 mean?

It is the nomenclature version. Version 3 is the colon-separated nomenclature in use since April 2010.

Can an HLA allele name change after it is assigned?

Yes. A name can gain a field when a close relative is named, gain or lose an expression suffix, move to a different family, or be deleted. The permanent HLA ID behind it does not change.

Where can I see which names changed in a release?

Allelelist_history.txt and Deleted_alleles.txt in the IPD-IMGT/HLA repository record every change. GET /v1/releases/diff summarises the differences between any two releases.

Sources

HLA-Verify is a research-and-evaluation tool for nomenclature and reference-release validation. It is not a medical device and this guide is not clinical advice. Send allele names, never patient identifiers.

Converting pre-2010 HLA allele names to current nomenclatureHow to convert colon-less HLA names like A*0101, Cw*0702, A*9201 and DPB1*0102 to current names, and the cases where adding colons gives the wrong answer.
HLA expression suffixes N, L, S, C, A and Q explainedWhat the letter at the end of an HLA allele name means, how IPD-IMGT/HLA assigns it, why it can change between releases, and the null-allele trap in matching.
How to read an HLA allele name, field by fieldWhat each part of a name like HLA-A*02:01:01:01 means, from the gene to the four fields and the expression suffix, with examples checked against IPD-IMGT/HLA.
Deleted HLA alleles: why names are withdrawn and how to trace themWhy IPD-IMGT/HLA deletes allele names, the reasons recorded in Deleted_alleles.txt, and how to trace a deleted name to its successor through the HLA ID.

All guides