Combined Paternity Index (CPI) Explained: What the Number on Your Paternity Report Actually Means

When your paternity test results arrive, you get more than a simple yes or no. You get a number called the Combined Paternity Index, or CPI, along with a separate figure called the Probability of Paternity. Most people glance at the "99.99%" line and stop reading. But the CPI is the value that actually drives that percentage, and understanding it can help you interpret an unusual result or explain the science to family or a court.

This guide walks through what the CPI number means, how the lab calculates it, why some tests produce values in the millions while others sit closer to a thousand, and what threshold your state may require if the results are used in a legal proceeding.

What the Combined Paternity Index Actually Is

The Combined Paternity Index is a likelihood ratio. It compares two possible explanations for the DNA patterns seen in the test:

• The tested man is the biological father of the child.
• A random, unrelated man from the same general population is the biological father.

A CPI of 10,000 means the DNA evidence is 10,000 times more likely if the tested man is the father than if a random unrelated man is the father. The number is not a percentage and not a probability on its own. It is a ratio of two probabilities.

That is an important distinction. A CPI of 500,000 sounds enormous, but it becomes meaningful only when you convert it into a Probability of Paternity, which is the figure most reports show at the top. We will get to that conversion shortly. First, it helps to see how the number is built one marker at a time.

How Each Genetic Marker Contributes: The Paternity Index

Modern paternity testing uses Short Tandem Repeat, or STR, analysis. The lab examines specific regions of DNA where a short pattern of letters repeats a variable number of times. At each of these regions, called a locus, a person carries two alleles, one inherited from the biological mother and one from the biological father.

For every locus tested, the lab compares the alleles in the child, the alleged father, and, if included, the mother. If the child carries an allele that could have come from the alleged father, the lab calculates a Paternity Index, or PI, for that specific marker. The PI is a mini likelihood ratio that answers one narrow question: at this single locus, how much more likely is the observed pattern if the tested man is the father than if a random man is?

The formula is straightforward. At a given locus, the PI equals X divided by Y, where X is the probability that the alleged father would pass the matching allele to the child, and Y is the frequency of that allele in the reference population. The rarer the shared allele, the smaller Y becomes, and the larger the PI grows.

That last point matters more than any other detail on the report. Rare alleles produce dramatically inflated PI values. Common alleles produce modest ones. This is why two paternity tests with the same overall conclusion can have wildly different CPI numbers.

Combining the Individual PIs Into the CPI

Once the lab has calculated a Paternity Index for every locus tested, the Combined Paternity Index is simply the product of all the individual PIs multiplied together. If a test examines 20 markers and produces PI values ranging from 1.2 to 8.5, the CPI is 1.2 x 3.4 x 2.1 x ... and so on through all 20 markers.

Because multiplication compounds rapidly, even modest per-marker PI values produce large CPI figures once you multiply enough of them together. Twenty markers each contributing a PI of just 3 would yield a CPI of over three billion. That is why the total can easily reach the millions or higher when every marker is consistent with paternity.

US Diagnostics Center analyzes up to 28 genetic markers per test, well above the industry baseline of 20 or more markers. More markers means more individual PI values multiplied together, which raises statistical confidence and helps distinguish between closely related potential fathers in cases where that might come up. You can read more about the underlying method on the paternity testing overview page.

Why Rare Alleles Change Everything

Population allele frequencies are the hidden lever in every PI calculation. Labs use published reference databases, often organized by ethnic background, to estimate how common each allele is in the broader population. Rare alleles are worth much more evidentially than common ones.

Consider a simplified example. If a child shares an allele with the alleged father that occurs in 40 percent of the reference population, then Y in the PI formula is 0.4, and the PI at that locus is roughly 1 divided by 0.4, or 2.5. Now imagine the shared allele occurs in only 2 percent of the population. Y becomes 0.02, and the PI jumps to 50. Same match, wildly different statistical weight.

This is why CPI values can vary so much between tests that all reach the same conclusion. Two children who each share every marker with their fathers might get CPI values of 50,000 and 5,000,000 depending entirely on how common or rare their inherited alleles happen to be. Both results mean the same thing in practical terms: paternity is essentially confirmed. But the raw number can look very different.

Another factor is ethnic reference database matching. Labs use population frequencies that match the tested family's background because allele frequencies differ across populations. Using the correct reference database is part of standard laboratory practice and is one of the quality control measures that accredited labs must document.

Converting CPI to Probability of Paternity

The Probability of Paternity is the figure most people focus on. It is expressed as a percentage, usually 99.99 percent or higher when paternity is confirmed. This number comes from the CPI through a formula rooted in Bayesian statistics:

Probability of Paternity = CPI / (CPI + 1) x 100

The formula uses what statisticians call a prior probability, which represents what you would assume before seeing any DNA evidence. Standard practice, endorsed by the International Society for Forensic Genetics, is to use a neutral 50/50 prior. That means the calculation starts by assuming the tested man is equally likely to be the father or not the father, and then the DNA evidence updates that assumption.

Under a 50/50 prior, the math works out cleanly. A CPI of 100 gives a Probability of Paternity of 99.01 percent. A CPI of 10,000 gives 99.99 percent. A CPI of 1,000,000 gives 99.9999 percent. The higher the CPI, the closer the probability approaches, but never actually reaches, 100 percent.

Some jurisdictions or expert witnesses use different priors, especially when other evidence in a case makes paternity more or less likely on its face. If a stronger prior is used, a lower CPI can still yield a very high probability. In practice, however, U.S. paternity reports almost always assume the neutral prior, and the CPI does the heavy lifting.

CPI Values You Are Likely to See on Your Report

Here is a reference table showing common CPI values, their corresponding Probability of Paternity under a neutral 50/50 prior, and the typical legal or interpretive context.

Notice the last row. When the DNA does not support paternity at multiple markers, the CPI is not simply low, it is zero. There is no gradient between "not the father" and "father." One exclusion at a single locus can sometimes be caused by a rare genetic mutation, so labs require exclusions at multiple markers before reporting a definitive exclusion.

Court Thresholds and Legal Standards

Most U.S. states have written into law a specific Probability of Paternity threshold that creates a legal presumption of paternity. If a paternity test result meets or exceeds that threshold, the tested man is presumed by the court to be the biological father unless other evidence contradicts the result.

The most common statutory thresholds are:

• 99% Probability of Paternity (CPI of at least 100) is the threshold set by many states, including Texas, Illinois, and several others. This is a rebuttable presumption, meaning it can be challenged with other evidence.
• 99.9% Probability of Paternity (CPI of at least 1,000) is required by some states as a stricter standard, particularly for cases involving child support enforcement or contested custody.
• Combined 100-to-1 requirement: A handful of states require both a CPI of 100 or greater and a Probability of Paternity of 99% or greater as parallel conditions.

Every state has slightly different rules, and the specific wording of the statute matters. For an authoritative summary of federal guidelines that influence state law in this area, see the federal Office of Child Support Services, which coordinates paternity establishment policy across states.

It is worth remembering that home paternity tests, sometimes called peace of mind tests, produce the same scientific results as legal tests. The difference is chain of custody. A legal test requires witnessed sample collection with verified identities, which makes the result admissible in court. Home tests skip that step, so the results are for personal knowledge only. If a court proceeding is possible, a chain-of-custody test is the right route from the start. The comparison between the two options is covered on the legal vs. peace of mind testing page.

The Role of CODIS Core Loci and Marker Selection

Not all genetic markers are created equal. The Combined DNA Index System, or CODIS, is a database maintained by the FBI that uses a core set of STR loci for forensic identification in the United States. In 2017, CODIS expanded from 13 to 20 core loci, and paternity testing labs generally include all of these plus additional markers for even stronger statistical resolution.

The NIST Short Tandem Repeat DNA Internet Database is the reference resource for allele frequencies and technical specifications of every STR marker used in forensic and relationship testing worldwide. Accredited labs use validated population databases derived from published, peer-reviewed sources so that PI calculations are consistent and defensible.

The specific markers chosen for a paternity test are selected for several qualities: high variability in the population, low mutation rates, no known association with medical conditions, and independent inheritance from one another. Testing 20 or more markers, or the up to 28 markers analyzed in US Diagnostics Center's home test panels, means the individual PI values are truly independent and can be multiplied together without statistical bias.

When the CPI Number Matters More Than a Simple Match

In straightforward cases where the tested man is the biological father, the CPI will be large, the Probability of Paternity will exceed 99.99 percent, and the result is unambiguous. Most reports fall into this category, and the specific size of the CPI is not something the family needs to think about further.

The CPI number becomes more important in a few less common situations:

• Related potential fathers. If the alleged father has a brother, father, son, or other close male relative who could also be the biological father, the CPI can help distinguish between them, especially when the mother's sample is included and additional markers are tested. Two brothers share about half their DNA, so their CPI values for the same child can both be positive but usually differ significantly.
• Mutations at a single locus. Occasionally a mother, father, and child triplet will show a mismatch at just one marker while matching at every other. This is often the result of a rare STR mutation during sperm or egg formation. A single-marker mismatch in an otherwise strong match is not automatically an exclusion, and the report will explain how the lab handled it.
• Missing the mother's sample. A motherless test still works, but the CPI will generally be lower than a test with all three samples, because the lab has less certainty about which allele in the child came from the mother versus the father. Some laboratories offer a mother DNA sample as an add-on available during checkout, and including it strengthens the statistical case.
• Legal proceedings with contested facts. When paternity is being challenged in court, opposing counsel may focus on the specific CPI value and the population database used. Understanding what the number means and how it was calculated helps you follow the technical arguments.

For anyone thinking about testing options beyond standard paternity, related relationship tests such as the full sibling DNA test at $139, the grandparent DNA test at $139, or the aunt/uncle DNA test at $139 use the same STR framework but produce a Combined Index that is interpreted differently based on the expected genetic relationship.

What to Look for on Your Own Report

When your paternity report arrives, here is a practical walkthrough of what to check.

1. The overall conclusion. The report will state either that the tested man is the biological father, that he is excluded as the biological father, or, in rare cases, that the result is inconclusive and additional testing is recommended.
2. The Probability of Paternity. This is the percentage figure. When paternity is confirmed by a reputable lab, this number will typically read 99.99 percent or higher. When the tested man is excluded, it reads 0 percent.
3. The Combined Paternity Index. This is the large ratio, often in the thousands, millions, or higher. A larger CPI means stronger statistical support for paternity. There is no maximum, and different tests will produce different CPI values depending on which specific alleles were inherited.
4. The individual marker table. Most reports include a table listing each locus tested, the alleles observed in the child and the alleged father, and the individual PI at that locus. This is where you can spot which markers contributed the most to the overall CPI, which alleles are shared, and whether any locus shows an unusual result.
5. The prior probability used. Reports usually state the assumed prior explicitly, most often 0.5, or 50/50. If a different prior is used, it should be labeled.
6. The lab's identifying information. Look for the lab name, address, and accreditation details. This information is what makes the report credible if you ever need to refer to it in a formal setting.

If any of these elements are missing or unclear, contact the lab. Reputable providers are used to walking customers through the numbers and will explain each section of your report.

US Diagnostics Center's home paternity test, priced at $79, produces the same statistical report as more expensive alternatives, with the same panel of up to 28 markers, the same PI and CPI calculations, and the same standardized interpretation. Lab processing takes 2 to 3 business days from sample receipt, and the total order-to-results window is 7 to 10 days with standard shipping. Kits and results are also available through the full home DNA test collection.

Frequently Asked Questions

What is a good CPI number on a paternity test?

Any CPI above 100 corresponds to a Probability of Paternity above 99 percent, which is the legal threshold in most U.S. states. Most accredited paternity labs will not report a confirmed paternity result unless the CPI is at least 100, and in practice the numbers are usually much higher, often in the tens of thousands or millions. There is no upper limit to the CPI, and a higher number simply reflects stronger statistical evidence, often driven by rare alleles or the inclusion of the mother's sample.

Can two different paternity tests give different CPI values?

Yes. Two tests of the same trio can produce different CPI values if the labs test different markers, use different population reference databases, or if the tests included different family members. The overall conclusion, father or not the father, will be the same across accredited labs. The specific CPI value can vary because it depends on the exact combination of markers analyzed and how common or rare the shared alleles are in the reference population used.

Is a Probability of Paternity of 99.99 percent as good as 100 percent?

In practical terms, yes. A Probability of Paternity mathematically cannot reach exactly 100 percent because the underlying calculation always leaves a tiny theoretical possibility that a random unrelated man could carry the same combination of alleles by chance. Once you reach 99.99 percent, that theoretical possibility is about 1 in 10,000. At 99.9999 percent, it is about 1 in a million. Any result above 99.99 percent is considered a scientifically confirmed match.

Do court-admissible paternity tests have a higher CPI than home tests?

No. The laboratory analysis is identical for both home tests and court-admissible legal tests. The same STR markers, the same PI calculations, the same CPI multiplication, and the same population databases are used. The only difference between the two types is chain of custody, meaning who witnesses the sample collection and how identities are verified. Chain of custody makes the result admissible in a courtroom but does not affect the mathematics of the CPI or the Probability of Paternity.

What happens if my CPI is between 0 and 100?

This is rare and usually indicates that some markers are consistent with paternity while others are not, or that the sample quality was borderline. A CPI in this range is generally considered inconclusive. The lab will typically request a recollection or additional testing before finalizing a report. A reputable provider will not report a definitive result based on a marginal CPI. If you receive an inconclusive result, contact the lab to review the individual marker data and discuss next steps, which often involve retesting with fresh samples or adding the mother's sample to the analysis. The how paternity testing works page covers what to expect from the process end to end.