The most accurate DNA test depends on what you’re testing for, but in purely technical terms, whole genome sequencing (WGS) at 30x depth or higher reads your entire genetic code and delivers the highest accuracy available to consumers. Standard consumer tests from companies like AncestryDNA and 23andMe use a different method called genotyping, which checks only a fraction of your DNA. Both approaches have strengths, but they are not equally precise, and “accuracy” means something very different when you’re exploring ancestry versus screening for health risks versus establishing paternity.
Genotyping vs. Whole Genome Sequencing
Most popular consumer DNA kits use genotyping arrays. These chips scan between 600,000 and 2 million specific points in your DNA, called SNPs, out of roughly 3 billion total base pairs. That means they’re reading less than 0.1% of your genome. For the positions they do check, genotyping arrays are quite reliable. But they leave enormous gaps, and the software fills those gaps through a statistical process called imputation, essentially guessing what the missing data looks like based on patterns found in reference populations.
Imputation introduces error. Research published in Frontiers in Genetics found that imputation error rates ranged from under 1% when the chip already captured a high density of SNPs, up to 18% when fewer markers were available. The size and genetic similarity of the reference population also mattered: error rates dropped from about 6.4% to 3.3% as the reference panel grew from 100 to 10,000 samples. In short, genotyping is accurate for what it directly measures, but the picture it fills in around those measurements can be unreliable.
Whole genome sequencing reads every base pair. At 30x depth, each position in your DNA is read approximately 30 times, which catches the vast majority of errors through sheer repetition. At 100x depth, each letter is read about 100 times, pushing accuracy even higher. WGS captures rare variants, structural changes, and regions of the genome that genotyping chips simply skip. For consumers who want the most complete and technically precise picture of their DNA, WGS is the gold standard. Companies like Nebula Genomics, Dante Labs, and Sequencing.com offer 30x WGS kits, with some providing 100x options at a premium.
Accuracy for Ancestry Testing
Ancestry results are where people most often feel misled by DNA tests, but the issue usually isn’t the raw data. It’s the interpretation. Every ancestry company compares your DNA to its own proprietary reference panels, collections of DNA from people with known regional backgrounds. Because these panels differ from company to company, your ethnicity estimates can shift noticeably depending on which service you use. One company might report 40% Scandinavian where another says 30%.
The genotyping itself is consistent across the major companies. Where they diverge is in how many reference populations they maintain, how finely they divide geographic regions, and how their algorithms handle overlapping genetic signatures between neighboring populations. For broad continental ancestry, all the major tests agree. For granular regional breakdowns, like distinguishing between Welsh and English heritage, results become less precise regardless of the platform. WGS won’t dramatically improve ancestry estimates because the bottleneck is the reference database, not the amount of DNA being read.
Accuracy for Health and Genetic Risk
Health-related DNA reports carry the highest stakes, and this is where accuracy gaps become genuinely dangerous. Consumer genotyping tests check only a handful of known variants associated with conditions like breast cancer, Alzheimer’s risk, or carrier status for inherited disorders. A clean result doesn’t mean you’re in the clear. It means the few variants the test checked came back negative.
More concerning is the false positive rate. A study published in JCO Precision Oncology examined patients who received consumer genetic screening and were then referred for confirmatory clinical testing. Researchers found that 40% of pathogenic variants identified in consumer raw data turned out to be false positives when retested in a clinical lab. For certain BRCA1/2 mutations (those outside the three founder mutations common in Ashkenazi Jewish populations), the false positive rate reached 69%. Many of these alarming results came not from the original test company’s reports but from third-party interpretation tools that consumers used to analyze their raw data files.
Clinical-grade genetic testing, ordered through a healthcare provider and processed by a certified lab, uses sequencing methods designed for diagnostic accuracy and includes professional interpretation. For health decisions, especially anything involving cancer risk or reproductive planning, clinical testing is the only version that should be trusted. Consumer tests are screening tools at best.
Accuracy for Paternity Testing
Paternity DNA tests work differently from ancestry or health kits. They compare specific genetic markers between a child and a potential father. The industry standard is 16 markers, though some tests analyze more for added confidence. Results are reported as a “probability of paternity,” which lands at either 0% (excluded as the father) or 99.9% (not excluded, and the statistical likelihood of parentage is overwhelming).
For legal purposes, such as child support or custody proceedings, the test must be collected under a chain-of-custody protocol, meaning a neutral third party witnesses the cheek swab collection and verifies the identity of each person tested. Home paternity kits use the same DNA analysis, but because no one verifies who actually provided the sample, they aren’t admissible in court. The underlying science is identical in both cases. The 99.9% probability threshold has been the accepted legal standard for decades.
Forensic DNA Profiling
Forensic DNA testing, the kind used in criminal investigations, operates on a completely different scale of precision. Since 2017, the FBI’s CODIS database has required DNA profiles based on 20 STR loci, short repeating sequences scattered across the genome. These loci have enough variation that the probability of two unrelated people matching across all 20 is astronomically small, often cited as less than one in a trillion for common populations. This makes forensic DNA the single most accurate form of human identification available. It is not available as a consumer product, but it sets the upper bound for what DNA testing can achieve in terms of individual identification.
What Makes a Lab Trustworthy
Regardless of the type of test, the laboratory processing your sample matters as much as the technology. In the U.S., any lab performing clinical testing must hold CLIA certification, a federal standard that sets minimum quality requirements. Labs can go further by earning accreditation from the College of American Pathologists (CAP), which conducts on-site inspections every two years using peer teams of trained professionals. CAP standards commonly exceed federal CLIA requirements, and their checklists are updated annually to reflect current best practices, with input from more than 500 pathologists.
When evaluating any DNA testing company, check whether their lab holds both CLIA certification and CAP accreditation. Most major consumer companies process samples in CLIA-certified facilities, but not all use CAP-accredited labs. For health-related results you plan to act on, CAP accreditation is the higher bar and the one worth looking for.
Choosing the Right Test for Your Goal
- Ancestry and ethnicity: Any major consumer genotyping kit (AncestryDNA, 23andMe, MyHeritage) will give you broadly reliable results. Differences between companies reflect their reference databases, not the quality of the DNA read. Testing with more than one company and comparing results is common.
- Health risk screening: Consumer kits offer a limited snapshot. If you have a family history of cancer, heart disease, or a genetic condition, clinical-grade sequencing ordered through a provider is far more accurate and comprehensive. Never make medical decisions based solely on a consumer test or third-party raw data interpretation.
- Maximum genetic detail: Whole genome sequencing at 30x or higher gives you the most complete and precise data. It’s more expensive (typically $200 to $600 for 30x, and higher for 100x) but captures variants that genotyping misses entirely.
- Paternity: A test analyzing 16 or more markers from a CLIA-certified lab will produce a definitive result. If you need legal admissibility, choose a provider that offers chain-of-custody collection.

