STR vs SNP Markers: Why Paternity Labs Use STR and Not SNP

If you have ever used an at-home ancestry kit and then shopped for a paternity test, you have probably noticed that the two tests cost about the same but produce completely different kinds of results. One tells you about your great-great-grandparents' ethnic origins. The other tells you with over 99.99% certainty whether a specific man fathered a specific child.

The reason is not price or marketing. It is the type of DNA markers each test analyzes. Ancestry kits use SNP markers. Paternity labs use STR markers. These two approaches look at completely different features of the genome, and each is designed to answer a different question.

This guide covers what each marker type actually does, why ancestry-style SNP testing cannot confirm paternity, and why STR analysis has been the gold standard for relationship testing for over 30 years.

What Are STR Markers?

STR stands for Short Tandem Repeat. These are stretches of DNA where a short sequence of letters, usually 2 to 6 letters long, repeats itself multiple times. For example, at one STR location, your DNA might read GATAGATAGATAGATAGATA. The sequence GATA has repeated 5 times. Someone else at the same location might have 7 repeats, or 11 repeats, or 15 repeats.

The number of repeats at each STR location varies widely between people. At any single location, there might be 10 or more possible repeat counts in the human population. Because each person inherits one copy of each STR from their mother and one from their father, you end up with a two-number signature at each location. For example, a person might read (12, 15) at one marker, meaning they have 12 repeats from one parent and 15 from the other.

Multiply that pattern across 20 or more STR locations and you get a highly individual DNA profile. The chance that two unrelated people share the exact same STR profile across all those locations is vanishingly small, typically in the range of one in several hundred billion.

What Are SNP Markers?

SNP stands for Single Nucleotide Polymorphism. These are single-letter variations in the DNA sequence at a specific position. At a given SNP location, most people might have the letter A while a smaller group has the letter G. SNPs are much simpler than STRs. Instead of a repeat count that varies widely, each SNP is essentially a yes-or-no question about which letter is present.

SNPs are useful because there are tens of millions of them scattered across the human genome. By checking a few hundred thousand SNPs at once, an ancestry company can compare your pattern against reference populations and estimate roughly where in the world your DNA signature best matches. This is how ancestry kits estimate ethnic composition.

The information per SNP is low. Any one SNP location gives you very little information about who you are, because so many people share the same letter. The power comes from aggregating hundreds of thousands of them.

Why Paternity Labs Use STR and Not SNP

The core difference is information per location. An STR marker with 10 possible repeat counts gives the lab roughly 10 times more discriminating information than a SNP with 2 possible letters. For a direct parent-child comparison, 23 STR markers is enough to produce probabilities above 99.9999% for a true inclusion and clear 0% for a true exclusion.

Here are the specific reasons STRs win for paternity testing:

• High discrimination per marker. One STR marker is more informative than hundreds of SNPs. A small panel of STRs is sufficient for a conclusive paternity result.
• Mendelian inheritance is clean. A child's STR profile at each location must contain one repeat count from each biological parent. If a child shows a repeat count that cannot have come from either tested parent, that is an exclusion. The logic is strict and unambiguous.
• Legal and forensic standardization. STR panels have been the standard for forensic DNA identification since the 1990s. Courts, immigration agencies, and the Social Security Administration all recognize STR-based paternity reports.
• Mutation rates are known and modeled. STRs occasionally mutate, meaning a child's repeat count may differ by one step from both parents even in a true relationship. These rates are well characterized, and labs account for them in the statistical model.
• Smaller sample needed. STR analysis requires very little DNA. A cheek swab from a child or adult provides more than enough material.

USDC's at-home paternity test analyzes up to 28 genetic markers. The standard panel is 23 autosomal STR markers, which produces probabilities at or above 99.9999% for true inclusions and clear 0% exclusions. For kinship tests like grandparent, sibling, or aunt/uncle, the lab can add up to 5 more autosomal STR markers to resolve inconclusive cases. The deeper article on autosomal STR markers and why paternity tests rely on them covers the biology in more depth.

Why Ancestry SNP Kits Cannot Confirm Paternity

If SNPs can trace your ethnic origins, why can they not confirm whether a specific man is a child's father? The short answer is that SNP data is good at population-level patterns but weak at individual-level verification.

• Each SNP is binary. Confirming paternity requires precise individual identification. A yes-or-no vote at any one location leaves too much room for coincidence.
• SNP matching uses probability intervals. Ancestry kits that attempt to detect relatives report ranges like "likely 2nd to 4th cousin" because the SNP pattern alone cannot distinguish precisely between close generation steps. STR analysis for paternity returns a sharp inclusion or exclusion, not a range.
• Reference databases skew the result. SNP-based relative matching depends on who else has uploaded their DNA to the service. If you and your biological father are both in the same database, the service may detect the match. If one of you is not in the database, there is no match to find.
• Not accepted as legal proof. No court, immigration agency, or federal benefits program accepts a consumer ancestry kit result as proof of paternity. The standards for relationship testing published by AABB require STR-based analysis performed in an accredited lab.
• Collection chain is unverified. Even if a consumer SNP kit produced a probability estimate for paternity, the sample collection would not be witnessed or chain-of-custody tracked. For legal purposes, that disqualifies the result.

The Common Confusion: "I Did an Ancestry Test and It Says We're Related"

Ancestry services often include a feature that compares your DNA to other users and estimates your relationship. These estimates are useful for discovering previously unknown relatives, but they are not designed to confirm a specific relationship with legal or medical certainty.

A typical ancestry service might report that two users share 1,200 to 1,400 centimorgans of DNA and estimate the relationship as "half sibling, uncle, or grandparent." That range is too wide to distinguish between those options. For a definitive answer, STR-based relationship testing is required. The article on why 1,160 to 1,397 cM is ambiguous walks through this in more detail.

If you have an ancestry match that raises a specific family question, the appropriate next step is a targeted relationship test with the people in question.

When Does SNP Testing Have a Role?

SNP analysis is excellent for what it was designed to do: estimate population ancestry, screen for inherited disease risk variants, and discover unknown biological relatives in a shared database. It is not a replacement for STR testing in any of the following situations:

• Confirming paternity, maternity, or any specific first-degree relationship
• Confirming second-degree relationships like grandparent, sibling, or aunt/uncle
• Any case headed to court, immigration proceedings, or federal benefits determination
• Any situation where a clear inclusion or exclusion is needed, not a probability range

For those scenarios, a lab that runs STR-based analysis with proper reporting is the correct choice.

A Note on the Technology Evolution

Research labs are increasingly using whole-genome sequencing for relationship testing, which captures both STR and SNP information along with everything else. For consumer paternity testing, sequencing is overkill. The result quality is the same as a 23-marker STR panel for direct parent-child comparisons, and the cost is substantially higher. STR analysis will likely remain the standard for routine paternity testing for the foreseeable future because it is accurate, well understood, legally recognized, and economical.

For deeper reading on how STR analysis is standardized across accredited labs, the NIST Short Tandem Repeat DNA Database maintains the reference profiles and quality standards used by forensic and relationship testing labs worldwide.

Frequently Asked Questions

Can an ancestry DNA kit tell me who my father is?

Not directly. Ancestry kits can suggest potential biological relatives if those relatives have uploaded their own DNA to the same service, and they can estimate a relationship range. They cannot confirm a specific paternity claim with the certainty required for legal, medical, or peace-of-mind purposes.

How many STR markers does a reliable paternity test need?

The industry standard minimum is 20 autosomal STR markers. USDC's test analyzes up to 28 markers, with 23 as the standard panel and up to 5 additional markers available for kinship cases that need more statistical weight.

Are STR results more accurate than SNP results?

For paternity and close-relationship testing, yes. STR analysis provides a sharp inclusion or exclusion. SNP analysis for relatedness provides a probability range that may span multiple possible relationships.

Why do some labs advertise hundreds of thousands of markers?

Those labs are using SNP arrays, usually for ancestry or disease screening. For a direct paternity question, those markers do not add diagnostic value beyond what 20 to 28 STR markers already provide.

Is a court-admissible paternity test different from a home test?

The laboratory analysis is the same STR-based process in both cases. The difference is the collection: court-admissible tests require witnessed sample collection, verified identification, and documented chain of custody. The science is identical.

If I already have an ancestry report, can I use that for a legal paternity case?

No. Legal proceedings, immigration cases, and benefits determinations require STR-based analysis from an accredited laboratory with documented chain of custody. An ancestry report does not meet any of those requirements.