How to Choose a Genetic Test

Many people compare genetic tests by the number of indicators or markers. Yet few realise that what matters far more is which traits a test analyses, what its conclusions rest on and how clearly they are explained in the report. These three criteria determine whether a genetic test will help you make decisions or remain a set of numbers with no clear practical meaning.

The range of consumer genetic tests looks uniform only at first glance. The names are similar, the packaging promises a complete picture of health, and the price can differ several times over. To choose a test deliberately, it helps to understand what lies behind these promises: which technology the laboratory uses, where its conclusions about predispositions come from and how the company presents the result. Below we go through the criteria that affect a test’s value and the signals that should give you pause.

What a Test Measures (and What It Does Not)

A genetic test analyses DNA variants linked to particular traits or risks. It does not measure your state of health at the moment of testing, and it does not predict the future with absolute accuracy.

An informed choice begins with understanding the method. Most consumer tests use SNP genotyping: they read preselected single-nucleotide DNA variants with a microarray that covers between 500,000 and more than a million positions in the genome. This is a common genotyping technology in its own right, not a marketing figure. As an alternative to genotyping, next-generation sequencing (Next Generation Sequencing, NGS), which includes whole genome sequencing (WGS), reads all the “letters” of DNA in sequence and yields broader data, while costing considerably more (Hull et al., Circulation, 2025). In other words, depending on the chosen technology, the informativeness of genetic testing can differ substantially because of these technological differences in analysis.

The second thing worth understanding concerns the nature of the conclusions. The genetics of complex traits, including most chronic conditions and functional characteristics, describes probabilities rather than individual scenarios. This is precisely what most consumer genetic tests examine. The polygenic architecture of such traits is made up of the combined contribution of many variants at once, so a test estimates a given person’s genetic predisposition relative to a relevant population rather than establishing a future diagnosis. The discriminative ability of such estimates in the general population is still limited, so a genetic report is better treated as context than as a prognosis (Lewis & Vassos, Genome Med., 2020).

For how a genetic report is structured and what its scales mean, see What a PRS is, and what the numbers in your report mean.

Not All Genetic Tests Are About the Same Thing

The single label “genetic test” covers products with different purposes. Before comparing them, it is worth determining which category each belongs to.

• Ancestry tests estimate the regions your ancestors came from. They are not designed to draw conclusions about health.

• Preventive tests assess predispositions related to nutrition, metabolism, sleep and response to physical load. These are probabilistic estimates for prevention, not diagnoses.

• Clinical/diagnostic tests look for specific variants associated with hereditary diseases. They are performed in clinical laboratories, often by sequencing individual genes or by targeted PCR analysis, with confirmation and the involvement of a physician.

Confusing these categories is risky. A preventive test will not provide a diagnosis, and an ancestry test will say nothing about health predispositions and risks. The AHA scientific statement makes a separate point: if consumer test data are used in clinical decisions, potentially significant findings should be confirmed by clinical methods (Hull et al., Circulation, 2025).

Which Traits a Test Analyses and What Its Conclusions Rest On

The number of indicators is one of the first characteristics people notice when choosing a test. Yet it says the least about a test’s quality. What matters is not how many traits appear in the report, but how relevant they are to your goal and how reliably they are supported. Three hundred superficially described predispositions are worth less than a smaller set built on quality data with transparent interpretation.

To judge how well founded a test is, it helps to distinguish two dimensions of accuracy: analytical validity — whether the test correctly determines that a DNA variant is present — and practical validity — whether that variant is genuinely linked to the trait the report refers to. A test can be technically flawless and still rest on a poorly supported link between variant and trait, so both dimensions matter.

Clinical validity comes from scientific evidence. Associations between DNA variants and traits are identified in genome-wide association studies (GWAS), which compare genetic variants with traits across large samples. From these data a polygenic risk score (PRS) is built, aggregating the contribution of many variants into a single measure of predisposition (Slunecka et al., Hum. Genom., 2021).

Two further questions help tell a better-founded test from a weaker one.

1. Whether the interpretation models are updated, since GWAS databases are constantly expanding.

2. Which populations the estimates are built on. The accuracy of polygenic estimates declines as a person’s genetic distance from the sample the model was trained on increases (Ding et al., Nature, 2023).

Most current PRS models are built on samples of European ancestry. For the Ukrainian population this matters, because it belongs to the European genetic cluster, though the accuracy of the estimates may vary depending on the specific model.

Another marker of quality is how a test explains its scales. A percentile reflects how far your genetic predisposition differs from that of most people in a reference population, not the probability that you will develop a particular condition. A good report always states that the comparison is made relative to a reference population, not in absolute terms.

For how to choose a DNA test in line with your needs and health goals, see the article on choosing a Prism panel (article №6).

How to Understand the Results and Apply Them in Practice

An estimate of genetic predisposition has practical value only when the results can be interpreted correctly and used to make health decisions. A 2023 review showed that the predictive ability of polygenic estimates on their own is mostly modest, and combining them with clinical data improves it only moderately (Koch et al., J. Community Genet., 2023). So what matters is not the figure alone, but how it is interpreted in the genetic report.

Personalised recommendations should be tied to the specific genetic features identified during the analysis. If a recommendation such as “get more sleep” appears identically in the reports of people with different genetic profiles, it is more of a general lifestyle tip than the result of a personalised interpretation of DNA.

A clear genetic report that can be used in practice answers three questions: what the number means, what it is compared with and what to do about it. If the report gives a specific indicator, it is important to distinguish genetic predisposition from the actual measured level: a genetic predisposition to a lower level of vitamin D does not mean the person already has a deficiency on a blood test. Precise communication is critical here, so that probabilistic information is not taken as a verdict (Koch et al., J. Community Genet., 2023).

Who Performs the Analysis, and What Happens to Your Data

Clinical laboratories operate under state and industry accreditation standards, whereas some DTC companies do not follow them (Cleveland Clinic, 2025). Laboratory accreditation is the first filter for assessing a provider.

When evaluating a genetic testing service, pay attention to the following:

• Where the analysis is performed and whether the laboratory holds the appropriate accreditation.

• Who is responsible for the genetic analysis — whether the laboratory performing it has the relevant expertise and sufficient experience.

• How your data are stored, to what protection standards, and whether they are shared with third parties.

• Whether you can delete your data on request.

• How long the result takes to prepare.

So when comparing services, note whether the company openly states where the analysis is performed, to what standards user data are protected and how long results take to prepare.

Where Differences Between Tests Come From

One of the most common questions is why two tests give different results for the same person. The reason is not that one of them is wrong, but that they rely on different data and methods.

Differences arise at several levels. Different methods and different laboratory approaches read different sets of SNPs, so the raw data already differ. Estimates are built on different sets of GWAS/PRS that vary in the size and composition of their samples. The PRS-building methods themselves yield different accuracy across populations (Wang et al., Annu. Rev. Biomed. Data Sci., 2022). Accuracy also depends on a person’s genetic distance from the training sample (Ding et al., Nature, 2023). Interpretation algorithms and threshold values differ as well — and as a result two well-founded tests can place their emphasis differently.

Which Tests Not to Trust

A few phrasings in advertising reliably signal that something is off with the science.

• “Fully decodes your genome” — most consumer tests perform SNP genotyping, not whole-genome sequencing. If a company does not state plainly that it uses whole-genome analysis, such a claim is a serious overstatement.

• “Predicts all future diseases and diagnoses deficiencies” — genetics estimates probabilities; it does not determine the future.

• “Guarantees personal treatment” — no consumer test guarantees a treatment outcome.

• “Switches genes on or off” — a test only analyses DNA variants and changes nothing in them.

Correct communication speaks the language of statistics and scientific precision. If a company speaks the language of guarantees and categorical promises, that is a reason to be wary.

How to Match a Genetic Test to Your Goal

The choice begins not with the test but with the goal. If you are interested in ancestry, or you are thinking about prevention and lifestyle, the relevant products are the ones to look at. If there is a specific clinical question or a heavy family history, that is the domain of clinical genetics and a physician, not a consumer test. This is why the test should be chosen and interpreted by a specialist.

Once the category is clear, assess the test against the criteria above: which traits it covers and how well they are supported, how clearly the result is presented, whether the recommendations are tied to your own variants, and who is responsible for the analysis and your data. The number of indicators is secondary to the quality of interpretation here. In the Apixmed Prism catalogue you can compare panels by their purpose, their list of traits and their field of application. This helps you match your own goal to what each test offers and choose the option that fits.

The results of a genetic test are not a diagnosis and not a substitute for a doctor’s consultation. An Apixmed Prism report provides genetic context that complements the findings of medical examinations and helps you make decisions together with your doctor.

Sources

1. Hull, L. E., Aday, A. W., Bui, Q. M. et al. (2025). Direct-to-consumer genetic testing for cardiovascular disease: a scientific statement from the American Heart Association. Circulation, 151, e905–e917. https://doi.org/10.1161/CIR.0000000000001304

2. Slunecka, J. L., van der Zee, M. D., Beck, J. J. et al. (2021). Implementation and implications for polygenic risk scores in healthcare. Human Genomics, 15(1), 46. https://doi.org/10.1186/s40246-021-00339-y

3. Lewis, C. M., & Vassos, E. (2020). Polygenic risk scores: from research tools to clinical instruments. Genome Medicine, 12(1), 44. https://doi.org/10.1186/s13073-020-00742-5

4. Koch, S., Schmidtke, J., Krawczak, M., & Caliebe, A. (2023). Clinical utility of polygenic risk scores: a critical 2023 appraisal. Journal of Community Genetics, 14(5), 471–487. https://doi.org/10.1007/s12687-023-00645-z

5. Ding, Y., Hou, K., Xu, Z. et al. (2023). Polygenic scoring accuracy varies across the genetic ancestry continuum. Nature, 618(7966), 774781. https://doi.org/10.1038/s41586-023-06079-4

6. Wang, Y., Tsuo, K., Kanai, M., Neale, B. M., & Martin, A. R. (2022). Challenges and opportunities for developing more generalizable polygenic risk scores. Annual Review of Biomedical Data Science, 5, 293–320. https://doi.org/10.1146/annurev-biodatasci-111721-074830

7. Cleveland Clinic (2025). DNA Tests & Gene Testing. Health Library, Diagnostics & Testing. https://my.clevelandclinic.org/health/diagnostics/22916-dna-test