Tumor–normal versus tumor-only sequencing

In one sentence

Tumor–normal sequencing compares tumor DNA with a suitable normal sample from the same person to help attribute variants, while tumor-only sequencing must infer their origin without that paired comparison.

The intuition

If you find an edit in a photocopy, comparing it with that person's starting version helps establish whether the edit was already there. Comparing it only with a public template leaves more uncertainty. The analogy has limits: a normal specimen is another biological sample, with its own cells, measurement quality and possible contamination.

How it works

The tumor contains a mixture of malignant and nonmalignant cells. A matched normal is a separately collected specimen from the same person, selected to represent their constitutional DNA for the intended analysis. “Normal” describes its analytical role; it is not a guarantee that every cell or genomic position is uncomplicated.

Both samples are sequenced, aligned and assessed at relevant locations. Evidence in the normal can support a constitutional origin; adequate tumor evidence with no corresponding normal evidence can support a somatic origin under the assay's rules. Insufficient normal depth and coverage leaves an attribution gap. Sample identity, contamination and the selected normal tissue must also be checked.

Tumor-only pipelines use population databases, allele fractions and other models to distinguish possible origins. A rare constitutional variant may be absent from population databases. A familiar cancer-associated gene or hotspot does not by itself prove that a variant arose only in the tumor. Jones et al. directly examined the interpretive consequences of omitting the matched normal.

Why it matters in cancer

Origin matters for biological interpretation, possible inherited-risk questions and immune-target design. But an authentic somatic variant may be a passenger; a constitutional variant may still matter for cancer biology. Attribution and function are different questions.

Paired research sequencing is not automatically a validated hereditary-cancer test. An inherited-risk question requires the appropriate clinical workflow, consent, interpretation and confirmation when indicated. Likewise, “not present in this normal sample” should remain a scoped measurement claim, not a guarantee about every cell in the person.

Assay card

Field	What to retain
Measures and method	Compare tumor and matched-normal sequence support, or use a clearly labeled tumor-only inference workflow
Input and tissue cost	Tumor DNA plus a suitable normal DNA sample for the paired design; each extraction uses material, although normal sampling need not use another tumor biopsy
Output and units	Variant coordinates, tumor and normal support counts and allele fractions, attribution labels and uncertainty
Thresholds	Separate usable coverage and detection criteria for tumor and normal; absence in an inadequately measured normal is uninformative
Failure modes	Sample mismatch, contamination, limited normal coverage, tumor mixture and rare constitutional variants missed by database filtering
Limits	Origin does not establish functional damage, inheritance counseling conclusions or drug benefit
Validation context	Check the intended somatic or germline use, sample types, analysis rules and confirmation policy

Common confusions

Matched normal versus a healthy reference cohort: the latter is not this person's paired DNA.
Tumor-only versus automatically somatic: specimen location does not establish variant origin.
Variant fraction versus inheritance: copy changes and cellular mixtures can alter the fraction.

Try it

A fictional tumor-only test finds a rare repair-gene variant at 48% allele fraction. A database does not list it. Can you label it inherited or somatic?

Answer: Neither conclusion follows from those facts alone. Review copy state and the assay's attribution model; suitable normal or clinical germline testing can address the origin question. Function remains another question.

Germline versus somatic: biological origins.
Variant calling: how sequence evidence is evaluated.

Sources

Source check: October 9, 2026. Expert and learner review remain pending. Examples are fictional.