Why mutation lands like a verdict
If a report says you have a mutation, the word does extra work that the DNA did not ask it to do. In everyday English, mutation means damage. Radiation. Comic-book accidents. A body that went off-script.
In genetics the older technical meaning was narrower and calmer: a permanent change in DNA sequence. That change can be harmful, harmless, or occasionally helpful, depending on the gene, the environment, and whether one copy or two are involved. The public ear does not hear that range. It hears a verdict.
This is why mutation vs variant is not pedantry. It is a fight about what a one-letter difference is allowed to imply before anyone has looked at evidence.
NIH and ClinVar prefer variant because most changes do not cause disease
MedlinePlus now opens its explainer with gene variant, and says the same kind of change used to be called a gene mutation. The reason is stated without drama: changes in DNA do not always cause disease, so variant is the more accurate term.
ClinVar, the public archive of submitted relationships between variants and conditions, is built on that vocabulary. Labs deposit interpretations of variants. They do not deposit mutations as a moral category. The five familiar ACMG/AMP words, pathogenic through benign, are labels for variants in a stated context.
Variant is the parent set. It includes SNPs that millions of people share, rare missense changes of unknown meaning, and the small number of changes that really do break a gene. Mutation, in popular use, pretends those groups are the same. They are not.
- Variant: a difference from a stated reference sequence
- SNP: a common one-letter variant
- Mutation, popular use: a change assumed to be harmful
- Mutation, older technical use: any permanent sequence change
Germline versus somatic is where the change lives
A germline variant is present in the egg or sperm, or in the fertilized egg that becomes a person. It is in essentially every cell. It can be passed to children. This is the kind of variant a saliva kit is designed to see, because cheek cells and white blood cells still carry the inherited spelling.
A somatic variant arises later, in a particular lineage of body cells. It is not in the egg or sperm, so it is not passed to children. Cancers are full of somatic variants. A mole, a tumor, a patch of skin that saw too much sun: those are somatic stories. A spit kit is a poor instrument for them, because the spelling may exist only in the tissue that grew it.
People mix the two because both are changes in DNA. Inheritance is the difference that matters. Germline is the family. Somatic is the biography of a tissue. Treating a tumor sequence as if it were an ancestry file, or an ancestry file as if it were a tumor sequence, is a category error. Mutation vs variant does not settle that geography. You still have to ask where the change lives.
- Germline: heritable, present from conception, visible in a saliva sample
- Somatic: acquired in body cells, not passed to children
- Spit kit: built for germline SNPs, not for a tumor's private spelling
De novo, mosaicism, and what a tube of saliva can miss
Some variants are new in a child and absent from both parents' ordinary cells. Those de novo changes explain genetic conditions that appear without a family history. They are germline in the child. They were not a family heirloom.
Mosaicism is the in-between state. A change happens after conception, so only a fraction of cells carry it. If those cells include some egg or sperm cells, an unaffected parent can still have an affected child. If they do not, the change stays a private fact of one body.
A consumer chip is a blunt tool here. It samples common germline SNPs in whatever cells made it into the tube. It is not a mosaic workup. It is not a test of every cell lineage. Absence from the file is not proof that a tissue-specific change does not exist.
Pathogenic is a submitted label, not a sentence
When ClinVar prints Pathogenic, a lab or expert group has submitted that the variant can cause a condition in a stated zygosity and clinical context. That is a claim about the variant. It is not a diagnosis of the person who grepped the rsID, and it is not a prediction that disease will appear on a given date.
Review status sits next to the word. Stars measure consensus and method, not how severe a disease is. One submitter without criteria and a practice guideline can share the letters Pathogenic and almost nothing else. Conflicting interpretations are their own aggregate state. Flattening a conflict into a red badge is how a spreadsheet invents certainty.
A variant of uncertain significance, a VUS, is not pathogenic-lite. It means the evidence was not enough to classify. Consumer reports that treat VUS as a finding to act on are doing work the label refuses to do. Benign and likely benign are also submitted interpretations, not a gold star for the rest of your genome.
Common spelling, rare disease alleles, and the frequency trap
A change can be common and still appear in a disease paper. rs1800562 in HFE is a standard example: not ultra-rare in European-ancestry groups, and still discussed in hereditary hemochromatosis. Frequency is context. It is not a classification.
gnomAD exists to make that context public. It counts how often an alternate allele appears among sequenced people after quality filters. A high frequency argues against a variant being a fully penetrant cause of a severe, pediatric, dominant disease. It does not prove the variant does nothing. A vanishingly rare allele is not automatically causal either.
The trap is the single emoji. Pathogenic plus common, or benign plus rare, are both possible. Read ClinVar and gnomAD as two questions. Then remember that neither database has met you.
What consumer reports flatten
A health-report chip looks at a designed list of common sites and then prints a story. The story often uses mutation for anything with a ClinVar word, risk for anything with a GWAS hit, and normal for anything the chip did not measure. Those are three different mistakes stacked in one PDF.
Zygosity disappears. Recessive disease in a homozygote is not the same object as a heterozygous carrier. Penetrance disappears. Some pathogenic variants cause disease in nearly everyone who carries the relevant genotype. Others raise a probability and leave most carriers unaffected. Condition lists disappear. A variant can be interpreted for one disease and be irrelevant to another.
A 23andMe file is not a BRCA test, not because BRCA variants are imaginary, but because a chip does not sequence BRCA1 and BRCA2. Most pathogenic variants in those genes will never be on the array. Calling the missing rows mutations you do not have is the mutation-vs-variant error in its purest form: a word standing in for a measurement that was never made.
Read the public record without turning it into a sentence about you
The useful habit is split. Look up the variant as a public object. Ask what it is, how common it is, and what has been submitted about it. Then, separately, ask whether a local file even called the site. Missing is not wild type. Called is still not a diagnosis.
genome.sh is a late tool for that first job. It prints public annotations for an rsID or a gene symbol. It does not apply ACMG criteria to your spit. It does not know if you are a heterozygote until a file that stays on disk is matched. That limit is the point.
If a clinician is in the picture, bring the right assay, not a renamed mutation. Single-gene tests, panels, exomes, and genomes see different things. Mutation vs variant will not settle a care decision. Evidence will, including zygosity, frequency, and the condition named on the record. This page is not medical advice.
Questions
What is the difference between a mutation and a variant?
A variant is a difference from a reference DNA sequence. Mutation, in popular use, implies that the difference causes disease. NIH and ClinVar prefer variant because most changes do not.
Why did NIH stop saying gene mutation?
MedlinePlus still explains the old word, then replaces it. Changes in DNA do not always cause disease, so gene variant is the more accurate public term.
What is a germline mutation vs a somatic mutation?
A germline variant is heritable and present from conception. A somatic variant arises in body cells later and is not passed to children. Saliva kits are built to see germline SNPs, not a tumor's private changes.
Does pathogenic in ClinVar mean I have a disease?
No. Pathogenic is a submitted label about a variant in a stated context. It is not a diagnosis of a person, and it does not know your zygosity until a real assay calls the site.
Is a VUS a mutation I should worry about?
A variant of uncertain significance means the evidence was insufficient to classify benign or pathogenic. It is not a milder form of pathogenic, and it is not a finding to act on from a consumer chip file.
If a variant is common, can it still be pathogenic?
Yes. Frequency is context, not a classification. Some disease-associated alleles are not ultra-rare in particular ancestry groups. Read ClinVar and gnomAD as two questions.
Is every SNP a mutation?
A SNP is a common one-letter variant. Calling every SNP a mutation imports a harm that most SNPs do not carry. Start with variant, then ask what evidence exists.
Can I diagnose myself from a mutation report?
No. This explainer is not medical advice. Clinical care needs a clinician and an assay that can see the relevant variants, which a consumer chip often cannot.