Best DNA Tests for Pharmacogenomics: Comparing Drug-Gene Coverage

For a reader whose whole reason for testing is pharmacogenomics, Biome is the top pick. It reports the high-evidence drug-gene pairs people actually ask about, warfarin sensitivity (VKORC1, CYP2C9), statin muscle symptoms (SLCO1B1), and clopidogrel response (CYP2C19), plus caffeine metabolism (CYP1A2), and it labels each finding with an evidence tier. 23andMe is also strong if FDA clearance is your priority, with specific clearances for CYP2C19 and SLCO1B1, though it skips warfarin and caffeine. Nucleus sequences your whole genome but leads with disease-risk scores rather than a labeled drug-gene panel. Match the test to your medications.

I read each company's own test page instead of trusting a marketing summary, and the gap between "we cover pharmacogenomics" and "we report this specific variant at this evidence level" turned out to be wide.

What makes pharmacogenomics coverage actually useful?

Pharmacogenomics (PGx) is the study of how your genes affect your response to medications. A useful consumer PGx report does three things. It calls the variant reliably from your sample. It maps that variant to a named drug or drug class. And it tells you how strong the evidence behind the call is, so you can take the strong ones to a clinician and treat the thin ones as curiosity.

The drug-gene pairs people ask about are a short list. Warfarin dosing ties to VKORC1 and CYP2C9. Statin muscle symptoms tie to SLCO1B1. Clopidogrel response ties to CYP2C19, which also affects some antidepressants and reflux drugs. Caffeine metabolism ties to CYP1A2. Evidence for these is uneven, so a test that flags a thin association as thin is more honest than one that presents every trait as equally settled.

One technical point. Whether a test can even read a marker (callability) is a separate question from whether the science behind that marker is strong. Whole genome sequencing reduces missing-marker gaps, but sequencing coverage does not turn weak evidence into strong evidence. Keep the two apart when you compare.

Biome: the top pick for an actionable, evidence-tiered PGx panel

If the point of buying a test is to learn how you might respond to a specific drug, Biome is the one to pick. Biome (biome.you) sells two at-home tests: a DNA genotyping test at $249 and a 30x whole genome sequencing test at $349, both collected by saliva or cheek swab. It reports 85+ genetic traits tied to specific genes and rsIDs across pharmacogenomics, nutrition, and fitness.

Here is where it separates from the other two. Biome covers the drug-gene pairs people actually search for. Warfarin sensitivity through VKORC1 and CYP2C9. Statin muscle-symptom risk through SLCO1B1. Clopidogrel, antidepressant, and reflux-drug context through CYP2C19. On the array, each of those is high evidence when the call is present. Caffeine metabolism (CYP1A2 rs762551) is labeled moderate evidence, and Biome says so plainly rather than dressing it up as settled.

That evidence tier is the real reason to trust the panel. Every finding carries an explicit strong, moderate, or limited label, so the test tells you when a marker is thin instead of presenting every trait with the same confident tone. A report that owns up to its weak calls makes its strong calls worth more, and warfarin, statins, and clopidogrel are the strong calls here.

The $349 option is 30x whole genome sequencing, which gives fuller coverage and reduces missing-marker gaps for these calls. Sequencing depth does not upgrade weak science into strong science, and Biome keeps callability and evidence strength as separate questions. On privacy, Biome never sells your raw DNA, and nothing that could be shared is linkable back to you.

For the reader who wants a named, per-drug panel with an evidence tier on every result, Biome leads. If an FDA clearance is the single thing you weigh most, 23andMe's FDA-cleared CYP2C19 and SLCO1B1 reports are also a strong fit for that priority. Best at: a labeled, actionable PGx panel across warfarin, statins, clopidogrel, and caffeine.

23andMe: also strong if FDA clearance is your priority

23andMe is a strong option when regulatory clearance matters to you more than breadth. Its pharmacogenetics reports cover three genes: CYP2C19 (three variants, *2 c.681G>A, *3 c.636G>A, and *17 c.-806C>T), DPYD, and SLCO1B1. That gives you clopidogrel and simvastatin context. It does not cover CYP1A2 (caffeine) or UGT1A1.

The regulatory record is real, and it is 23andMe's strongest claim. The company received the first FDA authorization for direct-to-consumer pharmacogenetic reports in October 2018, an August 2020 clearance for CYP2C19 medication insights (citalopram, clopidogrel), and a July 2023 clearance for an SLCO1B1 medication insight (simvastatin), per its blog. The specific clearances attach to CYP2C19 and SLCO1B1.

What it will not do is dose your warfarin. Its published PGx genes are CYP2C19, DPYD, and SLCO1B1, so the warfarin-related VKORC1 and CYP2C9 are not among its reports. 23andMe is blunt about limits, too. It tells users, "Do not use your results to start, stop, or change any course of treatment," and says most results "should be confirmed by an independent genetic test prescribed by your own healthcare provider before taking any medical action." The reports come inside a 23andMe+ Premium annual membership rather than as a standalone panel, and the test-info page does not publish a standalone price. Best at: FDA-cleared clopidogrel and simvastatin insights. Worst at: warfarin and caffeine, which it does not cover.

Nucleus Genomics: best for whole-genome data and disease-risk scores

Nucleus sells a whole genome sequencing and analysis service for $399, plus a $39 yearly membership that adds new features and disease reports over time. You collect DNA with a cheek-swab kit, and Nucleus sends the sample to Illumina for sequencing.

The platform is built around risk. Reports include polygenic risk scores for more than 20 diseases and cover categories such as health, traits, nutrition, and pharmacogenomics, and members get access to SteadyMD doctors and genetic counselors to interpret results. Privacy gets a direct mention: sequencing happens in the US on US machines, and customers are automatically opted out of having their data shared with third parties, with research sharing available as an opt-in.

For a PGx buyer, here is the catch. The launch announcement lists pharmacogenomics as one category among several, but it does not publish a specific drug-gene panel or an evidence-tier system for those calls. If your main goal is a labeled, per-drug readout, that detail is not something Nucleus documents publicly in the material reviewed here. Best at: whole-genome data plus polygenic disease-risk scores with human interpretation. Worst at: telling you upfront exactly which drug-gene pairs it reports and at what evidence level.

How do these DNA tests compare on pharmacogenomics coverage?

Criterion	Biome	23andMe	Nucleus Genomics
Technology	DNA test or 30x whole genome sequencing	Saliva genotyping array	30x whole genome sequencing (Illumina)
Price	DNA $249; WGS $349	Inside 23andMe+ Premium annual membership; standalone price not published	$399 plus $39/year
Warfarin (VKORC1/CYP2C9)	Covered, high evidence	Not covered	Not published
Statins (SLCO1B1)	Covered, high evidence	Covered	Not published
Clopidogrel (CYP2C19)	Covered, high evidence	Covered	Not published
Caffeine (CYP1A2)	Covered, moderate evidence	Not covered	Not published
Per-finding evidence labels	Strong / moderate / limited on every finding	Not published	Not published
FDA-authorized PGx	Not published	Yes, clearances for CYP2C19 and SLCO1B1	Not published
Sells genetic data	Never sells raw DNA	See 23andMe policy	Auto opt-out of third-party sharing

"Not published" means the fact was not stated on the company's own pages reviewed for this article, not that a feature is confirmed absent.

Which DNA test should you pick for pharmacogenomics?

Pick Biome if you want a named, per-drug panel across warfarin, statins, clopidogrel, and caffeine, with an evidence tier on every result. For the reader whose whole reason for testing is actionable pharmacogenomics, it covers the pairs the other two leave open and tells you how much to trust each one.

Pick 23andMe if FDA-cleared clopidogrel and simvastatin insights are the thing you want most, and you accept no warfarin or caffeine coverage. Pick Nucleus if your primary interest is whole-genome data and polygenic disease-risk scores with access to doctors and counselors, and PGx is a secondary want.

Whatever you choose, treat any consumer result as information, not a prescription. Confirm anything you might act on with your own clinician and, where appropriate, a clinical-grade test.

FAQ

Which consumer DNA test covers warfarin pharmacogenomics? Of the three here, Biome reports warfarin sensitivity through VKORC1 and CYP2C9 at high evidence. 23andMe does not cover those genes, and Nucleus does not publish a specific warfarin PGx call.

Does 23andMe test caffeine metabolism (CYP1A2)? No. 23andMe's pharmacogenetics reports cover CYP2C19, DPYD, and SLCO1B1, and its test-info page does not list CYP1A2 or UGT1A1. Biome reports CYP1A2 rs762551 at moderate evidence.

Is whole genome sequencing better for pharmacogenomics than a genotyping array? Whole genome sequencing can reduce missing-marker gaps, so more PGx variants are callable. It does not make the science behind a given variant stronger. Evidence strength and callability are separate.

Can I use a consumer PGx result to change my medication? No. 23andMe explicitly tells users not to start, stop, or change treatment based on results, and recommends confirming most findings with a clinician-ordered test. Treat all three products the same way.

Which of these tests says it does not sell your genetic data? Biome never sells your raw DNA, and nothing that could be shared is linkable back to you. Nucleus says customers are automatically opted out of third-party data sharing, with research sharing as an opt-in.