Alleles and zygosity: versions are different from copies
In one sentence
An allele is a version of DNA at a particular location; zygosity describes whether the versions present at that location are the same or different.
The intuition
Imagine having two editions of the same instruction. The editions can be identical or differ at one word. “Which version?” and “How many copies?” are separate questions. The analogy has limits: a tumor can contain several cell populations with different copy numbers, so there may be more than a tidy pair of editions.
How it works
A locus is a genomic location. An allele can describe a single base or a larger sequence version at that location. In an ordinary two-copy autosomal region, on a chromosome other than a sex chromosome, one inherited copy comes from each parent.
If the versions at a locus are identical, the genotype is homozygous. If they differ, it is heterozygous. These labels describe sequence state; they do not themselves say “healthy” or “damaging.” A hemizygous region has only one copy in an otherwise diploid setting, where most chromosome sets are paired. NCI hemizygous.
In cancer, copy number can change through loss or gain. Two copies do not guarantee two different alleles. Loss of heterozygosity (LOH) means that the different versions previously present are no longer both retained in the relevant cells or region. It can occur with copy loss, or be copy-neutral, with two copies of one version instead of one of each. Allele-aware data can reveal differences hidden by total copy number. Ciani 2022.
Use these fictional states to keep the axes separate:
| State at one locus | Total copies | Versions present |
|---|---|---|
| A/B | 2 | Two different versions: heterozygous |
| A/A | 2 | One version in two copies: homozygous |
| A only after loss | 1 | One remaining copy: hemizygous |
| A/A after loss of B | 2 | Copy-neutral LOH relative to the original A/B state |
The letters are sequence labels, not functional labels. The same state could have different biological consequences at different loci.
Loss of allele diversity need not reduce the total copy count.
Why it matters in cancer
A tumor-suppressor gene may lose useful function through changes involving multiple copies. A variant plus LOH can support a hypothesis of disruption, but retain which allele was lost and whether the remaining allele is damaging. The DNA state does not independently establish RNA or protein function.
Phase describes which variants occur together on the same DNA copy. Two variants in one gene are not automatically one on each copy. Biallelic disruption, damage involving both alleles, is a stronger biological claim than “two variants found.”
How it is measured
Variant allele fraction is the fraction of informative sequencing reads supporting a variant. It is not zygosity by itself. Normal-cell mixture, copy number and tumor populations affect that fraction. Purity, the malignant-cell proportion, and ploidy, the chromosome-copy context, help interpret it. Carter 2012.
Common confusions
- Allele versus copy: a sequence version can exist in several copies.
- Homozygous versus damaging: identical versions need not be harmful.
- LOH versus deletion: LOH can be copy-neutral.
- High read fraction versus biallelic loss of function: mixture, copies, phase and consequence still matter.
Try it
A fictional matched normal sample is A/B. The malignant cells are inferred to be A/A with two total copies. Did the normal total copy count rule out an allele change?
Show the answer
No. This is a copy-neutral LOH state relative to A/B. The consequence depends on the actual sequence and function of A. Neither its label nor the copy count establishes drug sensitivity.
Explain it back
“Copy number tells me ______; allele state tells me ______.”
One answer: “how many copies are present; which sequence versions those copies carry.”
Takeaway
Track versions, copy count and biological consequence separately, especially in a mixed tumor sample.
Related concepts
Sources and scope
Source check: October 9, 2026. General allele reasoning; the letters and copy-state exercise are fictional. Expert and learner review remain pending.
- NHGRI: allele.
- NCI: hemizygous.
- Ciani and colleagues, published online 2021 / issue 2022: allele-specific genomic analysis and copy-neutral LOH.
- Carter and colleagues 2012: purity, ploidy and absolute allelic copy inference.