Guides › IPD-IMGT/HLA releases

IPD-IMGT/HLA files explained: Allelelist, history, deleted and status

Updated 2026-10-30 · allele examples checked against IPD-IMGT/HLA 3.66.0 on 2026-10-30

The IPD-IMGT/HLA Database publishes its allele lists as plain CSV files in the ANHIG/IMGTHLA GitHub repository. Four of them answer most naming questions: Allelelist.txt (what is named now), Allelelist_history.txt (what every allele was called in every release), Deleted_alleles.txt (what was withdrawn and why) and Allele_status.txt (how well each allele is documented).

This guide describes each file as it appears in release 3.66.0, dated 14 October 2026, with the columns, an example row and the questions it answers.

The common header

Since release 3.32.0, the repository README says, all list files are CSV with a header, and header lines start with #. The header names the file, the date, the version, the origin, the repository URL and the author. In 3.66.0 the version line reads # version: IPD-IMGT/HLA 3.66.0.

Skip lines starting with # before handing a file to a CSV reader, and read the version from the header rather than from the file's location. A file copied into a pipeline keeps its header; a folder name does not prove anything.

Allelelist.txt: what is named now

The README describes it as "a csv file listing all alleles named at the time of the latest release".

Release 3.66.0 lists 47,424 names. That total includes the non-HLA genes the database also names, so filter by gene if you need HLA only. Names are full length as assigned: some have two fields, some four, and suffixes are part of the name.

Use it to answer "is this exact name current?". Do not use it to answer "is this a valid two-field name?": A*02:01 is not a row, because no allele is named exactly that, but it is a valid prefix of 519 rows.

Allelelist_history.txt: every name in every release

The README: "a csv file detailing for each allele the official name used in each release of the database".

The first column is HLA_ID; every other column is a release, written without dots and newest first: 3660, 3650 and so on back to 1050. In release 3.66.0 that makes 112 columns, one ID column and 111 releases. A release in which the ID had no name holds NA.

A row reads as one allele's naming history. HLA02908 is A*02:01:14 in 3.00.0 to 3.03.0, A*02:01:14Q in 3.04.0 to 3.65.0 and A*02:01:14 again in 3.66.0. Before 3.00.0 it carried the colon-less name A*020114.

Use it to translate a stored name into the current release, to trace an allele back through renames, and to compare any two releases. The guide on comparing two releases has tested code. The file is large, because it carries every release; read it once and index it by name and by ID.

The README also describes an allelelist folder with one Allelelist.XXXX.txt per release. At the v3.66.0-alpha tag it holds 67 files, from Allelelist.300.txt to Allelelist.3660.txt; the number is the release without dots, so 3100 is 3.10.0.

Deleted_alleles.txt: what was withdrawn and why

The README: "a csv file detailing all deleted allele names, with reasons for the deletion. This list also includes details of any suffix changes."

After the standard header comes a # Deleted Alleles line, then three columns:

The description is free text. It often names the allele the sequence turned out to be, or the name it became, but that link is not a separate column, so parse it with care and prefer the HLA ID in the history file where both exist. The deleted alleles guide lists the kinds of reason.

Not every name change appears here. Alleles that only gained a field, such as A*01:284 becoming A*01:284:01 in 3.66.0, have no entry; the change is visible only in the history.

Allele_status.txt: how well each allele is documented

The README describes it as detailing "for each allele how many times it has been submitted, from how many cells, the unconfirmed/confirmed status of the allele, if the CDS is fully sequenced and if the allele is cDNa or gDNA sequence". The columns are Allele, Cells, Groups, Confirmed, Start, End, Partial and Type.

Across the 47,424 rows in 3.66.0, 13,857 alleles are Confirmed and 33,567 Unconfirmed; 30,315 have a Full coding sequence and 17,109 are Partial; 28,555 are genomic (gDNA) and 18,869 coding-only (cDNA). An unconfirmed or partial allele is still a correctly named allele. The file tells you how much evidence and sequence sits behind the name, which matters when you interpret a rare result.

release_version.txt and the rest

release_version.txt is described as "a plain text file which denotes the current release version". In 3.66.0 its content is the six-line header itself, and the version is the # version: line.

The README lists other top-level files you will meet:

• sversion_history.txt, the sequence version of each allele in each release;
• md5checksum.txt, checksums for the top-level files;
• Nomenclature_2009.txt, the pre-2010 nomenclature;
• FASTA, MSF, alignment and XML folders with the sequences themselves;
• the wmda folder, with hla_nom.txt, hla_nom_g.txt, hla_nom_p.txt, rel_dna_ser.txt and rel_ser_ser.txt, covered in the G and P group files guide.

Large files are zip-compressed from release 3.56.0. A separate testing branch carries upcoming format changes, at least four weeks ahead where possible; it contains no unreleased names and is not for clinical or diagnostic use.

Which file answers which question

Licence and citation

The README applies the Creative Commons Attribution-NoDerivs licence and asks users to cite the database papers it lists. It says distributing a modified version of the data requires permission first, and that the maintainers are opposed to mirroring and prefer links. Check the current README before building the files into a product.

GET /v1/allele/{name}/history answers the history question for one name in a single call, with the release named in the response.

Frequently asked questions

Where is the list of all current HLA alleles?

Allelelist.txt in the ANHIG/IMGTHLA repository lists every allele named at the latest release, with its allele ID. Release 3.66.0 lists 47,424 names, including non-HLA genes.

How do I find an HLA allele's name in an older release?

Look up its HLA ID in Allelelist_history.txt and read the column for that release, written without dots (for example 3500 for 3.50.0).

What does Unconfirmed mean in Allele_status.txt?

It is the file's confirmation status for the allele, alongside how many cells it was reported from. An unconfirmed allele is still officially named.

Does Deleted_alleles.txt list every renamed allele?

No. It lists deletions and suffix changes, but alleles that only gained a field appear only in Allelelist_history.txt.