Which personal genomics service should I use in 2026 — and what to do given the 23andMe bankruptcy situation?

Why this comparison matters in 2026

The consumer genomics landscape shifted dramatically in 2025–2026.

In March 2025, 23andMe filed Chapter 11 bankruptcy. The court-supervised sale of the company — including its database of roughly 15 million customer genomic profiles — was approved mid-2025 and is still being executed. For roughly a decade, 23andMe was the default answer when someone asked "where should I get my DNA tested?" In 2026, it isn't.

At the same time, several alternative platforms have matured: Nebula Genomics is now an established whole-genome-sequencing option, Sequencing.com runs a useful "DNA App Store" model, Dante Labs continues as an EU-based WGS provider, and a new architectural model — Dark Bio — published a draft whitepaper in May 2026 proposing genomic-data sovereignty through hardware you physically hold.

Clinical pharmacogenomic testing through services like Color Health and Genomind has also moved into direct-to-consumer channels, often at lower cost than consumer genomics services when you only need PGx-relevant data.

This article compares the realistic options as of May 2026, with a deliberate lens toward what matters for peptide therapy and medication-management decisions specifically. For broader peptide-PGx context, see the pharmacogenomics cornerstone and the decision flowchart on whether to test before peptides.

Evidence tier: 2 — service-comparison information based on publicly verifiable product offerings, pricing, and operational status as of May 2026. The clinical-utility recommendations are based on CPIC guidelines (Tier 2) and the Swen 2023 Lancet RCT (Tier 1). Service-specific operational details are accurate as of publication but the landscape moves; revisit before purchase.

What you're actually choosing between — the four dimensions

When picking a personal genomics service, four dimensions matter more than the rest. The right service depends on which dimension you weight most heavily.

1. Coverage depth. SNP-chip arrays (what 23andMe used historically) read ~600,000 to 1 million selected positions. Whole-genome sequencing (WGS) reads all ~3 billion positions. WGS catches rare variants that chips miss — clinically important for comprehensive pharmacogenomics, family-history-of-disease workups, and rare-variant analysis. For most consumer purposes, SNP-chip is enough; for serious clinical use, WGS or clinical sequencing is better.

2. Clinical actionability. A consumer genomics report with broad ancestry / health-risk reporting is not the same thing as a CPIC-compliant pharmacogenomic interpretation. If you're testing for medication-management reasons, clinical PGx panels are more clinically actionable than consumer genomics services, even if the underlying genotype data overlaps significantly.

3. Data ownership and portability. Can you download your raw data? Can you delete your account and have the data actually deleted? Is the data sold or shared with research partners by default? Is the company likely to be acquired, breached, or change its terms over time?

4. Cost. Consumer SNP-chip $99–$199, consumer WGS $300–$600, clinical PGx panel $150–$500, comprehensive clinical sequencing $500–$2,000. Insurance partially reimburses clinical PGx in many cases; consumer services are out-of-pocket.

Different services optimize for different combinations of these dimensions.

Service-by-service breakdown

23andMe

Status as of May 2026: Operating, but in post-bankruptcy / post-sale flux. Recommend not submitting new samples until data-ownership questions are settled.

The dominant consumer genomics company from roughly 2013 to 2024. SNP-chip technology, broad ancestry and health-risk reporting, large database. The Chapter 11 bankruptcy in March 2025 was driven by financial losses and the December 2023 security breach that exposed genetic and ancestry information for roughly 7 million users through compromised accounts.

The court-supervised sale of 23andMe in mid-2025 included the customer database as a central asset. Post-sale terms of service are still being clarified. For new submissions in 2026, the risk is that the data ends up under terms different from what existed at submission, with no path back.

Existing 23andMe users can download their raw genotype data through the standard export tool (Account Settings → Browse Raw Data → Download Raw Data) and run it through third-party interpretation services independently of 23andMe corporate. The downloaded data is yours regardless of what happens to the company. Doing this now, before any further changes, is a sensible precaution.

Coverage: SNP-chip (~700,000 positions) Cost (historical, before bankruptcy): $99–$199 Best for: Existing users downloading their raw data for use elsewhere. Not recommended for new submissions in 2026.

Nebula Genomics

Status as of May 2026: Operating. The consensus pick for whole-genome consumer genomics in 2026.

Whole-genome sequencing at consumer pricing, 30x average depth (the standard for clinical-grade WGS). Returns raw VCF data you own. Has historically had a less aggressive monetization-of-data posture than 23andMe — privacy and ownership model is structured around your retaining custody of the raw data rather than around the company holding it as an asset.

The interpretation layer is reasonable for a consumer service: ancestry, health-trait reports, weekly variant-of-interest updates as new research publishes. For PGx specifically, the interpretation is informative but not CPIC-compliant in the clinical sense — for medication-management decisions you'd want a clinical pharmacist's interpretation of the underlying raw data.

Coverage: WGS, ~30x depth Cost in 2026: $299–$599 depending on subscription tier and promotional pricing Turnaround: 6–12 weeks typically Best for: Users who want complete genomic data they own, with the flexibility to re-analyze with third-party tools over time.

Dante Labs

Status as of May 2026: Operating. EU-based WGS competitor.

Italy-based whole-genome sequencing service. Ships internationally. Offers WGS at consumer pricing, often promotionally below Nebula's standard rates. Clear data-ownership terms; GDPR-framework privacy protections.

The trade-off: turnaround time has historically been longer than US-based services (often 3–6 months from sample to data, sometimes longer during promotional surges), and customer-service responsiveness varies by promotional cycle. If you're not in a hurry and price is the binding constraint, Dante is fair value. If you want faster turnaround and US-based support, Nebula is safer.

Coverage: WGS, 30x depth Cost in 2026: $199–$499 with frequent promotional pricing Turnaround: 3–6 months typical, longer during promotional surges Best for: Price-sensitive users who can wait. EU-based users particularly.

Sequencing.com

Status as of May 2026: Operating. The "DNA App Store" play.

Sequencing.com offers an interesting different model: upload raw genomic data (from any source — 23andMe, Nebula, Dante, AncestryDNA, clinical sequencing) and run individual analysis apps from third-party developers against it. The apps cover PGx, fitness, nutrition, ancestry, disease risk, and many specialty areas.

The model is curated rather than fully open — Sequencing.com reviews app submissions before listing — so trust still rests partly in the marketplace's review process. The contrast with Dark Bio's architectural model (where the sandbox boundary lets anyone publish an analysis safely) is real.

Sequencing.com also offers its own WGS at consumer pricing if you want their data + app store together. Reasonable middle ground if you want flexibility.

Coverage: Their own WGS, plus app store works with any uploaded raw data Cost in 2026: $0 to upload existing data; $99–$499 for WGS service; individual apps free to $99+ Turnaround: Apps run in days to weeks against uploaded data; WGS turnaround standard for the format Best for: People with existing raw data who want à-la-carte interpretation. People who want flexibility in interpretation rather than one company's bundled reports.

Dark Bio

Status as of May 2026: Whitepaper published May 2026; not yet shipping at consumer scale. The sovereignty-first architectural play.

Dark Bio proposes a different architectural model: your genomic data lives on a small dedicated device (the "Ark") that you physically possess, encrypted with two keys (one hardware-bound to the device, one held on your phone). Analysis apps are brought to the device and run inside a sandbox with no network access — the analysis sees your data, computes a result, and the result is the only thing released, with your explicit per-analysis approval.

The model is designed to make the kind of central-database breach that happened to 23andMe structurally impossible: there is no central database. Each Ark is one person's data on one device they hold.

This is genuinely different from every other service in this comparison — not a difference of pricing or coverage, but of trust architecture. The model is also designed to enable an open ecosystem: because trust rests in the sandbox boundary rather than in app authorship, anyone can publish analyses, and the platform doesn't need to gatekeep what apps users can run.

The trade-off in May 2026: it's not shipping yet. The whitepaper is detailed and the architecture is well thought through; consumer-scale availability is the gap.

Coverage: Whatever you load onto the device — supports import of WGS, SNP-chip, wearable streams, blood panels, imaging Cost in 2026: Not yet published at consumer scale Turnaround: N/A — not yet shipping Best for: People for whom genomic-data sovereignty is the top priority and who can wait. Worth tracking for the medium-term future.

For our broader analysis of the Dark Bio model, including takeaways for the personal-health-analysis space generally, see our strategic write-up on Dark Bio (if published).

Color Health / Genomind / clinical PGx panels through your physician

Status as of May 2026: Operating. The pragmatic option for PGx-specific testing.

Not a consumer genomics service in the broad sense — these are clinical PGx-focused tests. Direct-to-consumer in most US states (Color Health) or through a physician (most others). Cover the core pharmacogenomic enzymes (CYP3A4, CYP3A5, CYP2D6, CYP2C19, CYP2C9, SLCO1B1, plus phase II enzymes and HLA-B*57:01 for severe-reaction risk).

CPIC-compliant reporting that any clinical pharmacist can interpret directly. Often partially covered by insurance when ordered for medication-management reasons.

If you're testing for medication-management decisions specifically — and especially before peptide therapy where the relevant question is usually CYP-related — these services are faster, cheaper, and more clinically actionable than consumer genomics. The trade-off is narrower coverage: you get PGx-relevant variants, not your whole genome.

Coverage: Targeted PGx loci (CYP enzymes, phase II enzymes, HLA-B57:01) Cost in 2026: $149 (Color Health) to $500+ (clinical orders through hospital labs); often partially reimbursed by insurance Turnaround: 2–3 weeks typically Best for:* Pre-medication decision support. Pre-peptide PGx in any of the five YES scenarios from our decision flowchart.

Comparison matrix

| Service | Coverage | Data ownership | 2026 cost | Turnaround | Best for | |---|---|---|---|---|---| | 23andMe | SNP-chip (~700K positions) | Unsettled post-sale | $99–$199 (historical) | 6–8 weeks | Existing users downloading raw data — don't submit new in 2026 | | Nebula Genomics | WGS (30x) | Strong | $299–$599 | 6–12 weeks | WGS with clean data-ownership posture | | Dante Labs | WGS (30x) | Strong (GDPR) | $199–$499 | 3–6 months | Price-sensitive WGS, EU users | | Sequencing.com | Their WGS or upload your raw data | Strong | $0 to upload, $99–$499 for WGS | Apps fast, WGS standard | À-la-carte interpretation, app-store flexibility | | Dark Bio | Whatever you load — WGS, SNP, etc. | Strongest (hardware custody) | Not yet shipping | N/A | Sovereignty-first users willing to wait | | Color Health / clinical PGx | Targeted PGx loci only | Strong | $149–$500 | 2–3 weeks | Medication-management decisions, pre-peptide testing |

Decision guide by use case

"I want to know my pharmacogenomic status before starting peptides or any chronic medication." → Order a clinical CYP panel through Color Health or your doctor. Skip consumer genomics entirely. Cheapest, fastest, most clinically actionable path. See should I get a PGx panel before peptides for the full decision framework.

"I want my complete genome data, once, for ongoing use as new interpretation tools emerge." → Nebula Genomics is the default 2026 pick. Dante Labs if price is the binding constraint and you can wait. Both return raw VCF data you own.

"I have raw data from 23andMe / Nebula / AncestryDNA already and want to run specific analyses against it." → Sequencing.com's app store. Upload once, run apps as questions arise.

"I have a 23andMe account from before the bankruptcy and want to protect my data." → Download your raw data (Account Settings → Browse Raw Data → Download Raw Data) now. Then run it through Sequencing.com or specialized PGx tools as needed. The downloaded data is yours regardless of what happens to 23andMe corporate.

"Genomic-data sovereignty is my top priority. I'd rather wait than send my data to a cloud service." → Track Dark Bio for medium-term consumer availability. In the interim, if you must test, use Dante Labs or Nebula and keep the data under your local control.

"I have a family history of Wilson's disease and need ATP7B testing before any copper peptide protocol." → Clinical genetics service (Color Health, GeneDx, Invitae) ordered through a physician with genetic counseling. Consumer genomics is not the right path for inherited-disease testing where the result drives major medical decisions. See Wilson's + GHK-Cu contraindication for the operational detail.

"I have a family history of BRCA / cancer-risk variants and want comprehensive screening." → Same as above — clinical genetics service through a physician with genetic counseling. Not a job for consumer DNA testing.

What to do with your raw data once you have it

This depends on what you have and what you want from it.

SNP-chip data (23andMe, AncestryDNA): - Promethease (paid one-time, ~$12) for broad literature-based variant interpretation - Sequencing.com app store for specific specialty apps - Specialized PGx tools (StrateGene, etc.) for medication-focused interpretation - Clinical pharmacist consultation if you want CPIC-aligned dose recommendations

Whole-genome data (Nebula, Dante, clinical sequencing): - Same tools as above, with much richer underlying data - Variant interpretation services like Genoox or specialty PGx services - For peptide therapy specifically — clinical pharmacist interpretation against CPIC guidelines is the gold standard

Always: keep a local copy of the raw data on storage you control. If a service goes bankrupt (as 23andMe did) or changes its terms, your local copy is the asset you actually own.

Privacy and data ownership in 2026

The 23andMe situation made one point inescapable: a genomic database in a company's possession is an asset that can be sold, breached, or subpoenaed. The defense against this happening to you is not better policy at the company holding your data — it's not having your data held by anyone but yourself.

Three practical takeaways:

1. Download and locally back up your raw data from any service you use. If the company changes hands or fails, the data you've already downloaded is unaffected. 2. Read the terms of service. Look specifically for whether data is sold or shared with research partners by default (opt-in vs opt-out), what happens to your data on account deletion (actually deleted vs flagged-inactive), and what happens in an acquisition or bankruptcy. 3. Be cautious about adding data to an existing account. Each new layer of data — uploaded blood panels, integrated wearable streams, family-tree connections — increases the value of the database to a potential acquirer and the harm of a potential breach.

The Dark Bio architectural model (data on a device you hold, not in any company's cloud) is the cleanest answer to all three concerns. It is also not shipping at scale yet. For 2026, the pragmatic answer is to use services that give you the data and let you maintain a local copy, while being aware of the underlying custody problem.

What we don't know

Evidence tier: 5 — open questions.

• How the post-23andMe-bankruptcy data-ownership question will resolve. The court-supervised sale process is still being executed; the eventual outcome for existing customer data is not finalized.
• Whether and when Dark Bio will ship at consumer scale, and at what price point.
• Whether other custody-first architectural models (similar to Dark Bio) will emerge from competitors.
• Whether AI-agent interpretation of personal genomic data — the Sequencing.com-style app-store model extended with AI — will become the dominant interaction pattern.
• How insurance reimbursement for PGx panels will evolve; broader coverage would shift the calculus significantly.

Limitations

This article is service-comparison information as of May 2026; the landscape moves.

• Pricing changes frequently. Verify current pricing on each service's website before purchasing.
• Operational status changes. 23andMe's status is the most obvious example; other services have had pivots historically. Check current status before ordering.
• Regulatory environment varies by jurisdiction. US-based recommendations may not transfer cleanly to EU, UK, Canada, or other markets.
• Insurance coverage for clinical PGx varies by plan and indication. Confirm coverage before testing if cost matters.
• For inherited-disease testing, consumer genomics is not the right path — use clinical genetics services with appropriate genetic counseling.
• Marko Maal, MSc Pharmacy reviewed this article. Reviewer attribution does not constitute a doctor-patient relationship.

The bottom line

In 2026 there is no single best personal genomics service — the right pick depends on what you actually want from the test.

For pharmacogenomic information before peptide therapy or any chronic medication, skip consumer genomics and order a clinical CYP panel through Color Health or your doctor. Fastest, cheapest, most clinically actionable path. $150–$500.

For broad whole-genome data you own and can re-analyze over time, Nebula Genomics is the consensus 2026 pick. Dante Labs is the price-sensitive alternative.

For app-store-style à-la-carte interpretation against raw data you already have, Sequencing.com.

For inherited-disease testing (Wilson's, BRCA, etc.), use clinical genetics services with genetic counseling, not consumer DNA testing.

For medium-term sovereignty-first architecture, track Dark Bio.

For new submissions in 2026, do not use 23andMe until the post-bankruptcy data-ownership situation is settled. Existing 23andMe users should download their raw data now and run it through third-party tools independently.

What we'll be tracking

• Post-23andMe-sale data-ownership resolution
• Dark Bio consumer launch
• Insurance reimbursement policy changes for clinical PGx
• New custody-first architectural models from competitors
• Any consumer genomics service breach, bankruptcy, or terms-of-service change worth flagging

Related on this site

• Pharmacogenomics and peptide therapy (2026) cornerstone — broader PGx context
• Should I get a PGx panel before starting peptides? Decision flowchart — when testing is worth doing
• CYP3A4 and oral GLP-1 (Rybelsus) — operational detail for the CYP3A4-specific use case
• Wilson's disease and GHK-Cu contraindication — the ATP7B / clinical-genetics use case
• Finnrick vendor testing database — independent vendor verification for any peptide you purchase after the testing decisions
• Dark Bio whitepaper — primary source for the Dark Bio architectural model

References

• 23andMe. 2025. 23andMe Initiates Voluntary Chapter 11 Process to Maximize Stakeholder Value Through Court-Supervised Sale Process. https://investors.23andme.com/news-releases/news-release-details/23andme-initiates-voluntary-chapter-11-process-maximize — primary-source filing.
• Lavin Olivo H, et al. June 2025. Judge OKs sale of 23andMe — and its trove of DNA data — to a nonprofit led by its founder. NPR. https://www.npr.org/2025/06/30/nx-s1-5451398/23andme-sale-approved-dna-data — court-supervised sale outcome.
• TechCrunch. December 2023. 23andMe confirms hackers stole ancestry data on 6.9 million users. https://techcrunch.com/2023/12/04/23andme-confirms-hackers-stole-ancestry-data-on-6-9-million-users/ — the 2023 breach that contributed to the bankruptcy trajectory.
• Szilágyi P. 2026. Dark Bio: An Open Platform for Personal Health Analysis. Draft whitepaper, May 2026. https://dark.bio/whitepaper.pdf — primary source for the Dark Bio architectural model.
• Swen JJ, van der Wouden CH, Manson LE, et al. 2023. A 12-gene pharmacogenetic panel to prevent adverse drug reactions: an open-label, multicentre, controlled, cluster-randomised crossover implementation study. The Lancet. 401(10374):347–356. PMID 36739136 — clinical-PGx-panel evidence base.
• Clinical Pharmacogenetics Implementation Consortium. CPIC Guidelines. https://cpicpgx.org/guidelines/ — canonical source for clinical PGx dosing recommendations.
• US Federal Trade Commission. 2025–2026 statements on consumer genomic data privacy. https://www.ftc.gov — regulatory context for the post-23andMe environment.