The three categories of DNA services
Not all DNA services do the same thing. Understanding the categories helps you compare apples to apples:
1. Testing companies (collect and test your DNA)
These companies send you a kit, collect a saliva sample, genotype it on a chip, and give you reports. Examples: 23andMe, AncestryDNA, MyHeritage DNA, Living DNA.
What they do well: Collect your sample, generate raw data, and provide basic ancestry and/or health reports.
Limitations: Their health reports typically cover a small fraction of the actionable information in your raw data. They're designed for mass-market appeal, not deep health analysis.
2. Interpretation services (analyze your existing raw data)
These services take the raw data file you already have from a testing company and run deeper analysis. Examples: Your DNA Unlocked, Promethease, SelfDecode, Xcode Life, Genomelink.
What they do well: Extract more health information from data you already own. You don't need a new test.
Things to evaluate: What databases they reference, how they present results, whether they prioritize findings, and how actionable the output is.
3. Clinical genetic testing (medical-grade, doctor-ordered)
These are medical tests ordered by a healthcare provider, processed in CLIA-certified labs. Examples: Invitae, Color Health, Ambry Genetics, Natera.
What they do well: Highest accuracy, clinical-grade results, often with genetic counseling included. Results can directly inform medical decisions.
Limitations: Usually focused on specific conditions (cancer panels, carrier screening, prenatal testing). More expensive. Require a doctor's order in most cases.
What to look for in an interpretation service
If you already have raw DNA data and want health insights, here's what separates good interpretation services from mediocre ones:
What databases do they use?
The quality of genetic interpretation depends directly on the quality of reference databases. Look for:
- ClinVar — the NIH's clinical variant database with 341,000+ classified variants. This is the gold standard for disease variant interpretation.
- PharmGKB — Stanford's pharmacogenomics database with clinically annotated drug-gene interactions. Essential for medication-related findings.
- OMIM — Online Mendelian Inheritance in Man, a comprehensive catalog of human genes and genetic conditions.
- CPIC — Clinical Pharmacogenetics Implementation Consortium guidelines for drug-gene pairs.
If a service doesn't tell you where their data comes from, that's a red flag. Proprietary "algorithms" without transparent source databases are harder to validate.
How many variants do they analyze?
A curated database of 200-500 well-characterized health SNPs with detailed interpretations is more valuable than a dump of 10,000 variants with no context. Quality beats quantity, but the ability to screen against large clinical databases (like ClinVar's 341K+ variants) for disease risk is also important.
Do they separate high-confidence from speculative findings?
Good services clearly distinguish between:
- Pathogenic variants with strong evidence (needs attention)
- Well-studied common variants (useful context)
- Preliminary associations with limited evidence (interesting but don't act on it yet)
If everything is presented with the same level of urgency, you can't tell what actually matters.
Do they explain what to do with results?
Listing your variants without explaining what they mean or what to do about them is only marginally useful. Look for services that provide:
- Clear explanations in plain language
- Specific next steps (not just "consult your doctor" for everything)
- Priority levels so you know what matters most
- Context about when to involve a healthcare provider
Can you understand the report without a genetics degree?
If the output is a wall of rsIDs and p-values with no interpretation, it's built for researchers, not consumers. Good consumer tools translate genetics into language regular people can use and act on.
Red flags to watch for
- Claims to "diagnose" diseases from DNA alone. Consumer genetic testing is a screening tool, not a diagnostic one. Any service claiming to diagnose conditions from a raw data upload is overstating capabilities.
- Proprietary supplements sold alongside results. If a service analyzes your DNA and then conveniently recommends their own supplement line, the analysis may be biased toward selling products.
- No cited sources. Reputable services reference specific databases, studies, or guidelines. Vague claims without citations are marketing, not science.
- Urgent language designed to scare you. If every finding is presented as alarming or critical, the service is prioritizing engagement over accuracy.
- No privacy policy or vague data handling. If you can't find clear information about how your data is stored, shared, and deleted, don't upload your DNA there.
Questions to ask before choosing
- What format of raw data do they accept? Most accept 23andMe and AncestryDNA files. Some also accept data from other providers or whole genome sequencing.
- What does the output actually look like? Can you see a sample report before uploading? Does it separate findings by importance?
- What's the privacy policy? How is data stored, who has access, can you delete it, what happens if the company is sold?
- Is there ongoing support or just a one-time report? Genetic databases are updated regularly. Does the service re-analyze your data when ClinVar adds new classifications?
- What does it cost? Are there hidden charges? Some services offer a basic free analysis and charge for detailed reports. Others are one-time purchases. Know what you're paying for upfront.
The importance of updated databases
ClinVar reclassifies variants regularly. A variant classified as "uncertain significance" today might be reclassified as "pathogenic" or "benign" next quarter based on new research. This means:
- A one-time analysis gives you a snapshot, but genetic interpretation improves over time.
- Services that update their databases and re-analyze your data provide more value long-term.
- Re-uploading your raw data to updated tools periodically (even annually) can reveal new findings.
Matching service type to your goals
| Your goal | Best service type |
|---|---|
| Find out your ancestry | Testing company (23andMe, AncestryDNA) |
| Learn about drug metabolism | Interpretation service with PharmGKB coverage |
| Screen for disease risk variants | Interpretation service with ClinVar coverage |
| Confirm a specific pathogenic variant | Clinical genetic testing (Invitae, Color, etc.) |
| Carrier screening before pregnancy | Clinical genetic testing or interpretation service |
| Optimize nutrition and supplements | Interpretation service with nutrigenomics coverage |
| Cancer risk panel | Clinical genetic testing |
Frequently asked questions
Do I need a new DNA test, or can I use my existing raw data?
If you've already tested with 23andMe, AncestryDNA, or another consumer service, you almost certainly don't need a new test. Your existing raw data file can be uploaded to interpretation services for deeper health analysis. A new clinical test is only needed if you want to confirm a specific finding with medical-grade accuracy.
Are more expensive services necessarily better?
Not always. Price often reflects marketing budget and brand positioning more than analysis quality. Evaluate based on database coverage, report quality, and transparency — not price alone.
Can one service do everything?
Most services have strengths in certain areas. A pharmacogenomics-focused service may not have great ancestry tools, and vice versa. It's reasonable to use one service for ancestry and a different one for health interpretation.
How do I know if a finding is real vs. a false positive?
Check the ClinVar star rating, the inheritance pattern, and whether the variant is a true SNP (not an indel). For any finding that would change medical care, clinical confirmation with a CLIA-certified lab is recommended.