What Is a Genotype Test and How Does It Work?

A genotype test reads specific locations in your DNA to identify which genetic variants you carry. Rather than sequencing your entire genome from start to finish, most genotype tests zero in on known spots where people differ from one another, single letters of genetic code that have been swapped, or short stretches of DNA that repeat a variable number of times. The results can inform medical decisions about drug dosing, flag inherited disease risk, confirm biological relationships, or simply tell you about your ancestry. What makes genotype testing powerful, and occasionally confusing, is that the same basic technology serves wildly different purposes depending on which genetic markers are being read and who is doing the reading.

What a Genotype Test Actually Measures

Your genome contains roughly three billion pairs of chemical “letters.” The vast majority of those letters are identical from person to person. A genotype test focuses on the small fraction of positions where variation exists. The two most common types of variation that genotype tests target are single-nucleotide polymorphisms and short tandem repeats.

Single-nucleotide polymorphisms, usually called SNPs (pronounced “snips”), are spots where one DNA letter has been swapped for another. You might carry a C where someone else carries a T. Most SNPs have no obvious health effect on their own, but some influence how you metabolize a medication, how your immune system responds to a pathogen, or how likely you are to develop certain diseases. A microarray-based system can detect the vast majority of known SNPs in a single experiment by screening DNA against a chip loaded with complementary sequences.¹

Short tandem repeats, or STRs, are stretches where a tiny motif of one to six letters repeats back to back. The number of repeats at a given location varies from person to person, which makes STRs extremely useful for identification purposes. Over a million variable STR sites exist in the human genome, and variants at about 60 of those loci are known to cause genetic disorders, including Huntington disease and fragile X syndrome.² Because STRs mutate relatively quickly across generations and show high person-to-person diversity, they became the backbone of forensic DNA profiling and paternity testing.³

How a Sample Gets From Your Body to a Result

The process starts with collecting your DNA. Blood draws have long been the gold standard because blood yields large quantities of high-quality human DNA. But saliva-based collection has become increasingly common, especially in consumer and large-scale research settings, because it is painless and can be done at home. Saliva does come with a catch: a sizable fraction of the DNA in a spit sample belongs to bacteria and fungi living in your mouth. One comparison found that saliva samples contained an average of only about 37% human-amplifiable DNA, which brought genotyping call rates down to roughly 89% compared with about 99% from blood. When the human DNA fraction rose above about a third of the total, though, the call rate climbed to at least 96%, and the genotyping results matched blood-derived results with around 99% agreement.⁴

Studies in larger groups confirm that saliva works well enough for most purposes. In one cohort of over two thousand saliva specimens, about 6% failed quality thresholds entirely, usually because the DNA concentration was too low. Per milliliter of specimen, saliva actually yielded more total DNA than blood, but again, non-human DNA inflated those numbers.⁵ Separate work found that after purification, saliva DNA quality was comparable to blood DNA: 100% of blood samples and 93% of saliva samples could be successfully amplified by PCR.⁶ The practical upshot is that if you are doing a consumer ancestry test or a pharmacogenomics panel, a spit tube will almost certainly work. For high-stakes clinical or forensic work, blood or a buccal swab often remains preferred.

The Lab Techniques Behind Genotyping

Once DNA has been extracted and purified, the lab needs to “read” the variants of interest. Several technologies do this, and which one gets used depends on the question being asked and how many markers need to be checked.

PCR-based methods are the workhorse for situations where a lab needs to look at a relatively small number of known variants across many samples. The basic idea is to use short synthetic DNA sequences, called primers, that bind on either side of the region you care about, then copy that region millions of times until there is enough material to detect. Variations on this theme abound. One approach uses two pairs of primers designed so that each pair amplifies a product specific to one version of a variant, letting you see which versions a person carries in a single reaction tube.⁷ Real-time PCR methods such as TaqMan allelic discrimination assays add fluorescent probes that glow differently depending on which variant is present, making them especially practical for screening a handful of SNPs across hundreds of samples.⁸ In forensic settings, speed can matter enormously, and researchers have demonstrated direct PCR amplification of a multi-marker panel from a paper punch in under 25 minutes, with no separate DNA-extraction step needed.⁹

DNA microarrays, sometimes called gene chips, work on a different scale. A microarray is a small glass or silicon slide containing hundreds of thousands to millions of tiny spots, each loaded with a short DNA sequence complementary to a known variant. Your sample DNA, labeled with a fluorescent tag, washes over the chip. Where your DNA matches a spot, it sticks, and the fluorescence is recorded by a scanner. This lets a single chip read hundreds of thousands of SNPs in one pass. Early demonstrations showed that a single hybridization to a whole-genome microarray could detect more than 90% of the roughly 30,000 known SNPs between two yeast strains, along with insertions and deletions.¹ Consumer genetic testing kits from well-known companies rely on human versions of this technology, typically reading somewhere between 600,000 and over a million SNPs per sample.

Next-generation sequencing takes things further by actually reading the DNA sequence letter by letter, rather than just checking known positions. Whole-genome sequencing captures everything, while targeted panels focus on specific gene regions. In clinical oncology, for instance, next-generation sequencing can rapidly interrogate the cancer genome from a small tumor biopsy, identifying the mutations driving that particular patient’s cancer.¹⁰ Sequencing is more expensive and data-intensive than microarray genotyping, but it catches rare or novel variants that a chip designed to look at known positions would miss.

Clinical Uses of Genotype Testing

The clinical applications of genotyping have expanded rapidly, and three areas stand out: pharmacogenomics, carrier screening, and tumor profiling.

Pharmacogenomics

People metabolize drugs at different speeds, and much of that variation traces back to their genotype. The cytochrome P450 family of enzymes handles the breakdown of most clinically prescribed medications, and genetic variants in these enzymes can make you a fast metabolizer who burns through a drug before it works, or a slow metabolizer who accumulates dangerously high levels at a standard dose.¹¹ Pharmacogenomic genotype tests check a panel of variants in genes like CYP2D6, CYP2C19, and CYP3A5 to help clinicians choose the right drug and dose. One evaluation of pharmacogenomic testing found that the most commonly encountered enzyme phenotypes included CYP1A2 hyperinducers, CYP3A5 poor metabolizers, and CYP2B6 intermediate metabolizers.¹² This kind of testing is increasingly common in psychiatry, cardiology, and pain management, where the wrong dose of the wrong drug can have serious consequences.

Carrier Screening

Carrier screening tests look for variants that do not make you sick but could cause disease in your children if your partner carries a variant in the same gene. Expanded carrier screening panels now check for dozens or even hundreds of recessive conditions at once. In a large multi-center study of over 33,000 individuals, the overall combined carrier frequency was about 56% for roughly 200 autosomal genes, meaning more than half of people screened carried at least one recessive disease variant. Among families screened, about 5% turned out to be at-risk couples where both partners carried a variant in the same gene.¹³ A smaller study of 114 individuals found a carrier rate of about 45%, with most carriers harboring a single variant and a quarter carrying two.¹⁴ The numbers underscore why carrier screening is becoming routine in reproductive planning: being a carrier is common, but being an at-risk couple is rare.

Tumor Genotyping

In cancer care, genotyping the tumor itself has become a standard step for many types of cancer. The goal is to find the specific mutations driving that tumor’s growth so that targeted therapies can be matched to the molecular profile. Tumor genetic testing has become an important part of choosing appropriate therapies, and the number of mutations known to occur in targetable pathways keeps growing.¹⁵ Lung cancer is one of the best-studied examples, where multiplex genotyping and next-generation sequencing panels can identify actionable mutations from small biopsy specimens.¹⁰

Forensic DNA Profiling

Forensic genotyping is probably the form of genotype testing most people have heard of without realizing that is what it is. Crime labs extract DNA from evidence such as blood, hair, or saliva and amplify a standard set of STR markers. In the United States, the FBI’s CODIS database uses a core panel of STR loci to generate DNA profiles for identification purposes.¹⁶ The database contains profiles analyzed with panels of around 15 or more markers used for both biological-relationship testing and criminal identification.¹⁷

STR-based genotyping works well for straightforward identification, but things get trickier with complex family relationships. When the alleged father in a paternity case is a close relative of the true father, for example, the standard 20 CODIS STR markers still performed satisfactorily in most simulated scenarios, though researchers have explored supplementing them with additional STR and insertion-deletion markers for greater certainty.¹⁸ Some forensic labs also use SNP-based microarray systems that rely on a minisequencing step before hybridization to universal tag arrays, useful for degraded samples where the long DNA fragments needed for STR typing may not survive.¹⁹

Consumer Genotyping and Direct-to-Consumer Tests

Over the past decade, direct-to-consumer genotyping kits have brought what was once a clinical tool into the hands of millions of people. You order a kit, spit into a tube, mail it back, and receive reports on ancestry composition, trait predictions, and sometimes health risk. More than seven thousand genomes produced by six or more direct-to-consumer companies were made publicly available by consumers between 2011 and 2020, creating a growing body of data for researchers as well.²⁰

There is an important distinction between consumer tests and clinical-grade genotyping. In the U.S., health-related direct-to-consumer genetic tests are not recognized for diagnostic purposes. They are intended to provide genetic information associated with certain conditions, which may encourage you to discuss the results and their implications with a healthcare provider, but they are not a substitute for clinical testing ordered by a doctor.²¹ The chips used in consumer tests are designed for common variants in well-studied populations, and they can miss rare mutations or perform less accurately for people of underrepresented ancestries. If a consumer test flags something worrying, the standard advice is to confirm the finding with a clinical-grade test before making any medical decisions.

Interpreting Genotype Results

Getting a genotype result is one thing; knowing what it means for your health is another. For single-gene conditions, interpretation can be relatively straightforward: if you carry two copies of a known disease-causing variant, you have the condition. If you carry one copy, you are a carrier. But for most common diseases, the genetics are far messier. Heart disease, diabetes, and psychiatric conditions are influenced by thousands of genetic variants, each contributing a tiny nudge in one direction or another.

This is where polygenic risk scores come in. A polygenic risk score combines information across many genetic variants into a single number that estimates your overall genetic susceptibility to a given trait or disease. These scores are calculated using data from large genome-wide association studies and then applied to your personal genotype profile.²² The catch is that current polygenic risk scores typically explain only a small fraction of the total variation in a trait, even though genetics as a whole may be the single largest contributor to that variation.²³ A high polygenic risk score for heart disease does not mean you will definitely develop it, just as a low score does not guarantee you won’t. Lifestyle, environment, and luck all matter too.

Another interpretive challenge is the variant of uncertain significance, or VUS. When a lab sequences your DNA and finds a change that has not been definitively classified as either disease-causing or harmless, it gets labeled a VUS. The classification process relies on multiple lines of evidence, including whether the variant has been seen before in affected individuals, whether it is absent in healthy ones, and what its predicted effect on the resulting protein would be. When those lines of evidence are missing or contradictory, a confident call cannot be made, and the variant is reported as uncertain.²⁴ Receiving a VUS result can be frustrating because it essentially means “we found something, but we don’t know if it matters yet.” Over time, as more data accumulate, many VUS results get reclassified, but waiting for that reclassification is an uncomfortable reality of current genetic testing.

Accuracy, Validity, and What Can Go Wrong

When clinicians evaluate a genotype test, they think about it on two levels. The first is whether the test accurately reads the DNA variant it is designed to detect. The second is whether a positive or negative result actually corresponds to the clinical condition in question.²⁵ A test can be technically flawless at reading a SNP but clinically useless if that SNP turns out to have a weak link to the disease in question. Both dimensions need to hold up before a test is considered useful for guiding patient care.²⁶

Technical accuracy is generally high for well-established genotyping platforms, but it is not perfect. Different technologies can disagree on specific calls. When whole-genome sequencing data were compared head-to-head with targeted genotyping for pharmacogenomic markers in a pediatric cohort, six discordant genotypes were found across three genes, and hundreds of calls were missing or low-quality, mostly in one notoriously difficult gene called CYP2D6.²⁰ CYP2D6 is a gene that metabolizes a wide range of drugs, and its structure, full of duplications and deletions, makes it unusually hard to genotype accurately. That is an important caveat: the genes you most want to test are sometimes the hardest to get right.

At a larger scale, detecting subtle genetic events like mosaic copy-number changes, where only a fraction of your cells carry a deletion or duplication, requires specialized tools. One evaluation of a high-throughput detection method found a sensitivity of about 98% at a false-positive rate of roughly 34%, substantially outperforming older algorithms that achieved lower sensitivity with higher false-positive rates.²⁷ These are the kinds of technical tradeoffs happening behind the scenes. A test with very high sensitivity catches almost every true positive but lets in some false alarms, while a more conservative test might miss real events to keep the false-positive count low. Laboratories and clinicians choose their thresholds depending on the consequences of each kind of error.

When Genotype and Phenotype Do Not Line Up

One of the most common sources of confusion after a genotype test is the gap between what your DNA says and what your body actually does. The relationship between genotype and phenotype is not a simple readout. It is better understood as a connection between two differences: a change at the genetic level and a change at the observable level.²⁸ Carrying a risk variant does not automatically produce the associated condition, because other genes, environmental exposures, and random biological variation all modulate the outcome.

This disconnect catches people off guard in both directions. You might carry a variant associated with high cholesterol yet have perfectly normal levels because your diet and other genes compensate. Or you might have no known risk variants for a condition and develop it anyway. Genotype tests are probabilistic tools, not crystal balls. They shift the odds, sometimes substantially for single-gene conditions, but for complex traits they offer a partial picture at best. Understanding that limitation is arguably as important as understanding the test itself.