Guides › Nomenclature

HLA expression suffixes N, L, S, C, A and Q explained

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

Some HLA allele names end in a letter. That letter is an expression suffix, and it says the allele's protein is not expressed normally on the cell surface: N for null, and L, S, C, A or Q for other kinds of altered or uncertain expression.

The suffix is part of the name, not an annotation you can strip. Lose it and a non-expressed allele looks like an expressed one.

The six suffixes at a glance

According to the suffix page at hla.alleles.org, as of September 2026 no alleles have been named with the C or A suffixes. If you see one in a report, check it.

N: null alleles

N is used for alleles that have been shown, or are inferred, not to be expressed. The suffix page gives the common causes:

• A premature stop codon, from a point mutation or from an insertion or deletion that shifts the reading frame. A stop that falls early enough in the protein is inferred to leave nothing to anchor it in the membrane.
• A change to the starting methionine when there is no other methionine in the leader sequence to start from.
• A splice-site mutation that expression work has shown leads to no expressed protein.

Null alleles are not rare curiosities. The hla.alleles.org statistics for September 2026 list 509 null alleles at HLA-A, 413 at HLA-B and 451 at HLA-C.

Some nulls sit a long way from where you might expect. A*01:01:01:02N has a four-field name because its causative change is intronic: the protein is altered by splicing, but the mutation itself is outside the coding sequence, so it is named as an intronic variant of A*01:01:01.

L and S: expressed, but not normally

L marks an allele shown to have low cell-surface expression compared with normal levels. HLA-A*30:14L is the naming page's example, and HLA-A*24:02:01:02L shows that the change behind an L can sit outside the coding region.

S marks an allele whose protein is secreted as a soluble molecule and is not present on the cell surface. HLA-B*44:02:01:02S is the published example.

Both are expressed in some sense, which is exactly why they are flagged: an assumption of normal surface expression would be wrong.

Q: questionable expression

Q is the cautious suffix. The suffix page says it is used when a sequence "may or may not be expressed" and the work to decide has not been done. Typical triggers include:

• a non-synonymous change at amino acid 101 or 164 of a class I allele, which affects a disulphide bond;
• a small in-frame insertion or deletion;
• a premature stop inside the transmembrane domain;
• a change to the start codon when another methionine is available in exon 1;
• a mutation in the first two or last two bases of an intron, which can disturb splicing.

HLA-DQB1 is an exception for one splice variant: exon 5 is variably expressed there, and the known splice change before it does not affect expression, so those alleles do not get a Q.

Suffixes can change between releases

A suffix records the best evidence at the time. The suffix page says assignments may be revised as new data become available, and IPD-IMGT/HLA publishes a list of suffix changes with reasons.

When a suffix changes, the name changes. The history of one allele shows how:

The deletion note in the IPD-IMGT/HLA history reads: "Suffix added to allele becoming A*02:01:14Q (March 2011). Suffix then removed from allele becoming A*02:01:14 (August 2026)". A system that stored A*02:01:14Q in 2025 is holding a name that is no longer current, and a lookup against release 3.66.0 should say so and give the successor.

The null-allele trap in matching

A null allele can share its group, and even its exon 2 and 3 sequence, with an expressed allele. In release 3.66.0, A*24:09N belongs to the G group A*24:02:01G along with the expressed A*24:02 alleles. The 2010 nomenclature update's own G-group example, A*02:01:01G, includes the null alleles A*02:43N and A*02:83N.

So a typing that resolves only to the G group cannot by itself rule a null allele in or out. Likewise, a typing reported only as A*24 covers both A*24:02 and A*24:09N, although one is expressed and the other is not.

When the HLA-Verify match endpoint compares a recipient carrying A*24:02 with a donor carrying A*24:09N at an 8/8 framework, it counts the A locus as a mismatch and raises a null_allele flag, rather than counting two names in the same group as a match. The 8/8 and 10/10 guide walks through that example.

How to handle suffixes in software

A few rules prevent most suffix bugs:

1. Never strip a trailing letter to "clean" a name. N, L, S, C, A and Q are part of the name. The letters G and P are different: they mark group names, covered in G groups and P groups.
2. Do not parse the suffix as a fifth field. It attaches to the last field, whatever that is: A*24:09N has two fields, B*44:02:01:02S has four.
3. Store the release with the name. A suffix can be added or removed in a later release, as A*02:01:14 shows.
4. Treat a null as a distinct allele in comparisons. Two-field reduction of A*24:09N keeps the N; the API's two-field form of A*24:09N is still A*24:09N.

Frequently asked questions

What does N mean at the end of an HLA allele?

It marks a null allele: one that has been shown or is inferred not to be expressed, most often because of a premature stop codon. A*24:09N is the example on the hla.alleles.org naming page.

Are there any alleles with the C or A suffix?

Not as of September 2026, according to hla.alleles.org. The suffixes are defined in the nomenclature but no allele currently carries them.

Can an allele lose its suffix?

Yes. Suffixes follow the evidence and can be revised. A*02:01:14 gained a Q in March 2011 and lost it again in August 2026, becoming A*02:01:14 in release 3.66.0.

Is a Q allele expressed?

It is unknown. Q means the allele carries a change that may affect expression but expression has not been confirmed either way.