Every home paternity kit on the market prints a big accuracy number on the box. Usually it is 99.99% or higher. What almost none of them explain is where that number comes from, and how it changes when the lab reads more or fewer markers on your DNA sample. The number is not a marketing figure. It is a direct output of how many locations on the genome the lab compared between the child and the alleged father. The specific marker count the lab uses changes the math.
This piece walks through what a marker actually is, why more markers produce stronger results, why the industry has landed on 20 as a baseline and higher-end labs run 23 or up to 28, and when the extra markers actually move the needle. It is written for buyers trying to compare kits, not for population geneticists — but the underlying statistics are laid out plainly enough to explain the box.
What "marker" means in a DNA paternity test
A genetic marker in the context of relationship testing is a specific location on the genome where the DNA sequence varies enormously from person to person. The NIH National Human Genome Research Institute glossary defines a marker as an identifiable DNA sequence used to track a chromosome or a location within it. In paternity testing, the markers used are almost always short tandem repeats, abbreviated STRs. An STR is a short DNA sequence, usually 2 to 6 base pairs long, that repeats itself several times in a row at a fixed position on a chromosome. What varies between individuals is how many times the sequence repeats at that specific location.
For example, at a marker called D3S1358 on chromosome 3, the repeating unit is the four-base sequence "TCTA." Some people have it repeated 12 times in a row at that location. Others have it repeated 18 times. The number of repeats is inherited from your parents, one count from each. That two-number combination — say 14, 17 — is your genotype at that marker.
The lab reads the same marker from the child's DNA and from the alleged father's DNA. If the child's two numbers are 14 and 20, and the alleged father is 14, 18, and the mother (when tested) is 17, 20, the child's 14 came from the alleged father and the 20 came from the mother. That is a match at that marker. Match enough markers and the statistical case becomes airtight. Fail to match even one, and the alleged father is excluded outright.
Why the number of markers matters so much
Each marker on its own is a coin flip with a lot of possible outcomes. At a single marker, a random unrelated man might happen to share a genotype with the child roughly 1 in 10 to 1 in 30 times, depending on how common that specific number of repeats is in the population. That is not enough to prove anything. Any given man has a real chance of matching by pure coincidence at one marker.
The moment you add a second independent marker, the probability of coincidental match multiplies. A 1-in-15 chance at marker A combined with a 1-in-20 chance at marker B becomes 1 in 300 for the pair. Add a third marker and it becomes 1 in several thousand. This multiplication is the entire foundation of statistical DNA identification. It is the same math that makes forensic DNA identification reliable, and it is documented in the reference data maintained by the NIST Short Tandem Repeat DNA Database, which catalogs allele frequencies for every marker commonly used in forensic and relationship testing.
By the time you get to 15 or 16 independent markers, the probability that a random unrelated man would match the child at every single one drops below one in one billion for most population groups. That is where the 99.99%+ probability of paternity numbers on paternity test reports come from. It is not a courtesy figure — it is a direct calculation from allele frequency data at each marker.
Where the industry standard of 20 markers came from
The FBI's Combined DNA Index System, known as CODIS, originally used 13 core STR markers for forensic identification. In 2017 the FBI expanded that core set to 20 markers to strengthen international database compatibility and reduce the already-small chance of coincidental matches in very large databases. That 20-marker core has become the practical floor for reputable relationship testing labs in the U.S. — testing fewer than 20 markers today is unusual and generally reflects an older or lower-cost workflow. The industry standard phrasing across accredited providers is "20 or more markers."
At 20 well-chosen independent STR markers, the probability of two unrelated people matching at every one of them by chance is astronomically small — far less than one in one trillion for most population groups. For a paternity test, where the biological father must share one allele at every marker with the child, the exclusion power is essentially total. If the alleged father is not the biological father, 20 markers will exclude him with certainty. If he is, 20 markers will produce a probability of paternity of 99.99% or higher.
Why some labs run 23, and why USDC runs up to 28
Even with 20 markers being statistically overwhelming in the average case, there are situations where more markers help. The most common one is population homogeneity. When the tested parties come from the same ethnic or ancestral background, they are more likely to share common alleles at any given marker. Not because they are related, but because the population as a whole draws from a smaller allele pool. In those cases, the effective information from any single marker is lower, and the combined statistical case is a little weaker for the same number of markers than it would be for a genetically diverse pair.
Testing more markers restores the statistical strength. If the population at one marker only has 6 common alleles in wide circulation, testing an additional marker that has 12 common alleles in circulation is worth more than adding another 6-allele marker. Labs that run 23 markers are usually adding higher-information markers on top of the CODIS 20, chosen specifically because they provide additional discrimination in mixed-population cases.
USDC's home paternity test can analyze up to 28 genetic markers. The standard panel is 23 autosomal STR markers, which produces probabilities of paternity at or above 99.9999% for true inclusions and unambiguous 0% exclusions for true exclusions. Those 23 handle the overwhelming majority of paternity cases with no ambiguity at all. The International Society for Forensic Genetics publishes the interpretation conventions most accredited labs follow, including the 50/50 prior probability standard used to calculate the final number on your report.
When the extra markers actually fire
The additional markers above 23 in USDC's system are not run on every case. Paternity testing does not need them, because 23 well-chosen markers already push the statistical case beyond the threshold at which more markers meaningfully change the conclusion.
The extra markers are reserved for kinship testing where the standard panel returns an inconclusive result. Kinship tests — grandparent, aunt or uncle, and sibling — are structurally harder than a direct parent-child comparison, because the tested relative shares only 25% of DNA with the child on average, not 50%. That halves the information content of every shared allele, and the combined statistical output is expressed as a Combined Relationship Index, or CRI, rather than a probability of paternity.
A CRI greater than 10 supports the claimed relationship. A CRI less than 0.1 does not support it. A CRI between 0.1 and 10 is the inconclusive zone, and it happens more often on kinship tests than it does on paternity tests — sometimes 5 to 15 percent of cases, depending on the specific relationship and the ethnicity of the tested parties. When a kinship test lands in the inconclusive zone, up to 5 additional markers are available to push the CRI out of the gray zone and into a definitive answer. Those markers are not needed on paternity tests because paternity tests are already definitive at 23 markers. The extra 5 are targeted specifically at the harder statistical problem kinship testing creates.
Why "more markers" is not automatically better
It is easy to assume that if 20 markers is good and 23 is better, then 40 or 50 must be even better. It does not work that way in practice.
Above a certain marker count, the added statistical certainty per marker drops off sharply. Going from 15 to 20 markers takes the probability of coincidental match from roughly one in a billion to one in a trillion. Going from 20 to 30 markers takes it from one in a trillion to one in a quadrillion. Both of those are numbers so far beyond human intuition that neither one changes the practical outcome. If the alleged father is included at 20 markers, he is included. If he is excluded at 20 markers, he is excluded.
What matters at higher marker counts is not headline accuracy but robustness against edge cases: unusual mutation events, degraded samples, mixed populations, or one specific marker not producing a clean read. In those situations, having 3 to 5 extra markers in reserve means a clean answer without asking the customer to recollect samples. It is insurance more than raw accuracy.
The other reason not to chase very high marker counts on every case is chemistry cost. Adding markers means more reagents, more instrument time, and more analyst review per sample. On the vast majority of paternity cases, that additional cost buys no additional information. Running 23 markers as a standard panel and reserving the extra 5 for the specific cases that need them keeps the cost model realistic without compromising any conclusion.
Comparing paternity test kits on marker count
If you are shopping for a home paternity test, the marker count is one of the clearest technical differences between kits, and it is one of the few numbers on the box that translates directly into a statistical outcome. A few practical points for reading a marker-count claim:
- Anything below 20 is worth questioning. If a kit specifies fewer than 20 markers, or does not specify at all, the workflow may be older or lower-cost, and the statistical margin is smaller than the industry standard.
- 20 or more is the industry baseline for reputable providers. A kit that specifies "20 or more markers" is meeting current accredited-lab standards. That is where the 99.99%+ probability of paternity numbers you see on report cards actually come from.
- 23 or higher adds robustness. Kits from higher-end labs that run 23 or more markers as a standard panel absorb population overlap and edge cases more comfortably. USDC's home paternity test is $79 and runs up to 28 markers total, with 23 as the standard panel.
- "Advanced marker analysis" is not a number. If a kit uses phrases like "advanced analysis," "premium markers," or "expanded panel" without printing the actual marker count, the marketing is doing the work the specification is not. Ask for the specific number in writing before ordering.
The reason marker count matters at the consumer level is not that 99.99% is meaningfully better than 99.9%. It is that the marker count tells you which era of lab science you are buying, and how much statistical margin the lab has to absorb an edge case before it needs to ask you to swab again.
What this means for kinship tests
For kinship tests, marker count matters more than it does for paternity, because the underlying statistical problem is harder. A grandparent DNA test, an aunt or uncle DNA test, and a sibling DNA test all cost $139 and start with the same 23-marker standard panel that USDC uses for paternity. When that panel returns a Combined Relationship Index in the inconclusive 0.1-to-10 zone, up to 5 additional markers can be run to push the answer to a definitive conclusion.
That is where the "up to 28" number in USDC's marketing comes from. It is not a claim that every test runs 28 markers. It is a description of the ceiling — the total number of unique autosomal markers the lab can bring to bear on a case when the standard panel is not enough. For paternity, 23 is almost always enough. For kinship, the reserve of 5 additional markers keeps a much higher percentage of cases from getting stuck in the inconclusive zone.
Frequently Asked Questions
Is a 20-marker paternity test accurate enough?
Yes. Twenty independent STR markers, chosen from the current CODIS core set or similar, produce a probability of coincidental match far below one in one trillion for most population groups. That is enough to identify the biological father with a probability of paternity of 99.99% or higher, and to exclude a non-father with certainty. Testing more markers adds robustness against edge cases but does not meaningfully change the accuracy on typical samples.
Why do some labs run 28 markers when 20 is statistically enough?
The extra markers are held in reserve for kinship testing where the standard panel returns an inconclusive Combined Relationship Index. Kinship tests are structurally harder than paternity because the tested relative only shares 25% of DNA with the child on average. When the CRI lands between 0.1 and 10 — the inconclusive zone — up to 5 additional markers can push the answer into a definitive result. On direct paternity tests, the extra markers are usually not needed.
Can a paternity test with more markers ever contradict one with fewer markers?
Not in a meaningful way. A paternity test either produces an inclusion (with a probability of paternity typically 99.99% or higher) or an exclusion (0%). Running more markers on the same DNA never turns an exclusion into an inclusion or the reverse. It can sharpen the specific probability figure — pushing 99.99% to 99.9999% — but the outcome is the same. Where more markers can help is on a borderline kinship case, where the extra data resolves an inconclusive standard panel into a clear answer.
How do I check how many markers a home paternity kit uses before I buy?
Read the product page or the kit box for a specific number. Reputable providers state the marker count explicitly. If the marketing uses phrases like "advanced marker analysis" or "premium panel" without a number, contact customer support and ask for the specific count in writing before ordering. Fewer than 20 markers is worth questioning; 20 or more meets the industry baseline; 23 or more is the higher-end tier.
Does USDC run all 28 markers on every paternity test?
No. USDC's standard paternity panel is 23 autosomal STR markers, which is already statistically overwhelming for a direct parent-child comparison. Up to 5 additional markers are available for kinship testing when the standard panel returns an inconclusive Combined Relationship Index and more statistical weight is needed to reach a definitive answer. The 28 figure is the ceiling of unique markers the lab can bring to a case, not the count run on every case.
0 comments