Why Two Home DNA Tests Can Give Different Results (And What to Do About It)

A parent orders a home paternity test from one company. The result says 99.99% probability of paternity. A few weeks later they order a second test from a different provider, expecting to confirm the answer. The second result is inconclusive, or worse, it points the opposite direction. Now the family is more confused than before they started.

This scenario is more common than most product pages will tell you, and it does not always mean either lab did something wrong. There are three reasons two home DNA tests on the same biological question can produce different answers. Understanding all three tells you which result to trust and what to do next.

Reason 1: The two labs ran different numbers of markers

DNA relationship testing is a statistics problem. The lab looks at short tandem repeat (STR) markers scattered across the autosomal chromosomes and calculates a likelihood ratio: how much more likely the observed overlap is if the two people are related as claimed versus if they are unrelated. More markers means a stronger, more discriminating calculation.

The industry has settled on a few standard panel sizes:

• 13 to 16 markers: the older CODIS core panel, still used by some budget providers and legacy pharmacy kits. Reliable for straightforward paternity trios, weaker for anything else.
• 20 markers: the modern CODIS baseline, adopted by the FBI in 2017 and now the minimum most reputable labs run.
• 23 autosomal STR markers: the higher-tier standard now common in home paternity products. This is USDC's standard panel.
• Up to 28 markers: a reserve of additional autosomal loci that a lab can bring online for kinship inconclusives, which is USDC's reflex protocol.

Two labs testing the same biological question can produce different answers when one runs 16 markers and the other runs 23. On a clean paternity trio (mother, child, alleged father), even 13 markers usually settles the question. On a kinship test (grandparent, sibling, aunt/uncle), the answer is much more sensitive to the number of markers analyzed. A 16-marker sibling test can land in the 0.1 to 10 inconclusive range while a 23-marker test on the same samples resolves cleanly. The FBI's overview of the current CODIS core loci is available at the FBI CODIS page.

If two of your results disagree and the labs used different panel sizes, the one with more markers is the more reliable answer.

Reason 2: The interpretation ranges are the same, but the scores are close to the borderline

Kinship tests (grandparent, sibling, aunt/uncle) report a Combined Relationship Index or Combined Sibling Index. The interpretation ranges are the same across reputable labs:

• Above 10: supports the claimed relationship.
• Between 0.1 and 10: inconclusive.
• Below 0.1: does not support the claimed relationship.

A borderline result at, say, 8 from one lab and 12 from another lab is not a real disagreement. Both labs are reporting the same underlying signal. One landed just below the "supportive" threshold and one landed just above. The random distribution of which alleles a child inherited from which parent means small differences on the same samples across two labs are expected in the borderline zone.

The way to tell if two results actually disagree is to look at the numeric index, not just the categorical verdict. If both labs report an index between 0.1 and 10, both labs are telling you the same thing: the current test cannot resolve the question. The right next step is not to keep re-testing at different labs, it is to add markers or add a relative. The International Society for Forensic Genetics publishes the framework labs across the world use for handling inconclusive kinship results.

Reason 3: Sample collection was different on the two tests

The DNA analysis is not the only variable. What the lab receives to analyze matters just as much as what they do with it. Two common sample-quality issues that can push a result across a threshold:

Contamination from food, drink, or another participant

If a cheek swab was collected within 30 minutes of eating, drinking, smoking, or brushing teeth, food residue and mouth bacteria can dilute the human DNA on the swab. If the two participants' swabs touched each other or a shared surface before sealing, a small amount of cross-contamination can shift the marker read. On the first test, sample technique might have been sloppy; on the second, careful. That alone can move a borderline result across the threshold.

DNA yield from the swab

A light, quick swab produces less DNA than a firm 30-second swab. The lab has ways to work with lower yield, but marginal samples are more likely to produce a marker read that gets flagged for repeat analysis or misses a marker entirely. If one lab got a solid sample and the other got a marginal sample, the marginal sample's result carries more uncertainty even if the report does not say so on the face of it.

The step-by-step technique that maximizes DNA yield is covered in the guide on how to collect a cheek swab that passes lab QC. If the two tests used different collection instructions or one participant collected differently the second time, that alone can account for the different result.

Which result should you trust?

In order, the questions to ask when two home DNA tests disagree:

1. Which lab ran more markers? More is more discriminating. If one used 13 markers and the other used 23, trust the 23-marker result.
2. Are both results in the borderline range? If both indexes fall between 0.1 and 10, the tests are not disagreeing; they are both telling you the current sample cannot resolve the question. Next step is add markers or add a relative, not re-test at a third lab.
3. Was sample collection different? If one participant followed the 30-minute rule and the other did not, or if one collection was cleaner than the other, the result from the cleaner collection is more reliable.
4. What does each lab say about their panel? A lab that names its panel size and marker count on its product page is easier to evaluate than one that just says "state-of-the-art DNA analysis." Vagueness about panel size is a signal to look elsewhere.

If you cannot answer any of those questions from the reports you have, the fastest path to a definitive answer is to submit fresh samples to a lab that runs at least 23 autosomal STR markers and has a documented reflex protocol for inconclusive results. That is what USDC's process is built for. Standard order-to-results, including shipping, is approximately 7 to 10 business days. Lab processing itself is 2 to 3 business days once samples arrive.

When to retest at USDC

If your prior result came from a pharmacy kit or a low-marker legacy lab, and the answer either surprised you or was borderline, retesting on a 23-marker panel with reflex up to 28 markers usually resolves the question. The relevant USDC products depending on the biological question:

• Home paternity test ($79): mother-child-alleged father trios, or father-child duos when the mother is not available.
• Home maternity test ($129): questions about biological motherhood.
• Home full sibling or home half sibling tests ($139): confirming shared biological parentage between two children.
• Home grandparent test ($139): closing a paternity gap when the alleged father is not available.
• Home aunt or uncle test ($139): same use case as grandparent test when a grandparent is not available.

For kinship tests specifically, adding the biological mother's sample when possible sharpens the result. On borderline first-pass results, USDC's protocol reflex-runs the additional autosomal markers on the same submitted sample before returning a final report. No new order or re-collection is required for the reflex step.

Frequently asked questions

Can two paternity tests on the same trio really give different answers?

On a clean mother-child-alleged father trio with clean samples, the answer is unambiguous even at 13 markers. If two tests on that scenario disagreed, sample quality is almost always the reason. If your prior result was on kit swabs collected without the 30-minute rule, a fresh set of samples collected cleanly is likely to resolve the disagreement.

Should I test at three different labs to break a tie?

No. Two labs disagreeing is a sample-quality or panel-size problem, not a "majority vote" problem. Sending the same borderline samples to a third lab will most likely produce another borderline result. The productive next step is to submit fresh samples to a lab with more markers, add a relative for a kinship test, or both.

Is a more expensive test more accurate?

Not necessarily. Price does not correlate with panel size or lab quality. A $79 test that runs 23 markers is more discriminating than a $150 test that runs 16. Ask about the panel size before ordering.

Does USDC re-run tests if I submit a previous result from another lab?

USDC will run a fresh test using its own 23-marker panel with reflex up to 28 markers if needed. USDC does not re-analyze another lab's raw data, because different labs use different marker panels and different quality controls. Fresh samples processed on a single validated workflow is the cleanest way to resolve a disagreement.

What if the borderline result is on a kinship test rather than a paternity test?

Kinship tests are more sensitive to marker count than paternity tests, so adding markers (as USDC's reflex protocol does) helps more here than on a paternity trio. If the reflex step still returns an inconclusive, adding a second relative in a follow-up test is the next step. The guide on kinship DNA testing when only one relative is available covers the ranking of which relatives close the gap fastest.