Sample type matters less than collection quality. Blood is the clinical standard and usually gives the cleanest DNA. Saliva is easier and gives good data when it is collected properly, but it contains oral bacteria and fails quality checks more often. Cheek swabs are gentle and cheap but often yield less human DNA than a good saliva sample, according to a saliva-kit maker's own material.
For ancestry or general curiosity, use whatever your kit supplies and follow its instructions exactly. For a result that will guide medical decisions, a clinician-ordered blood test is the safer default.
The three sample types side by side
| Blood | Saliva | Cheek swab | |
|---|---|---|---|
| How collected | Needle draw by a trained collector | Spit into a tube at home | Rub a swab inside the cheek |
| Convenience | Lowest: needs a visit or trained collector | High: no needle | High: quick and painless |
| Human DNA purity | Very high | Lower: includes oral bacteria | Variable, depends on technique |
| Typical use | Clinical and research sequencing | Consumer ancestry, health and some sequencing kits | Ancestry, paternity and some children's tests |
This table summarizes common patterns. Individual providers run their own extraction and quality-control processes, so ask yours which sample types it accepts and why.
What the research found
A 2020 study in BMC Medical Genomics sequenced the whole genomes of 531 blood samples and 502 saliva samples from a hospital biobank of cardiac patients and compared quality. Its main results:
- 46% of saliva DNA samples failed the sequencing quality-control check, compared with 6% of blood samples.
- About 10.7% of reads from saliva came from the oral microbiome, against 0.09% for blood. Software can filter out reads that do not match the human reference, so this mostly costs sequencing efficiency.
- In five directly paired samples, more than 95% of single-letter variants found in blood were also found in saliva. Agreement was lower for rare variants (about 90%) and for copy-number variants (about 76%).
The authors concluded that saliva DNA that passes strict quality control can be acceptable when blood is unavailable, but is less reliable for copy-number changes. That is a single cohort with its own collection conditions, so it does not give a failure rate for your kit. It does show where saliva is weakest, which links to what we say about structural variants and CNVs.
A quality-control failure means the lab could not extract enough clean human DNA to reach its coverage target. It describes the sample, not your genome, and it is far more common with saliva than with blood. If you are sequencing at depth, ask the provider how it handles low-yield samples before you buy.
Why kits fail, and how to avoid it
Most failed samples come down to a few things you control:
- Not following the pre-collection rules. Kit instructions commonly ask you to avoid eating, drinking, smoking or chewing gum for a set time before spitting. Food residue and fluids dilute and contaminate the sample.
- Too little sample. Saliva kits need a specific volume, usually marked on the tube. Foam and bubbles do not count.
- Skipping the stabilizer step. Many tubes release a preserving liquid when you close the cap. Follow the order of steps printed in your kit.
- Heat and delay. Post the kit promptly and keep it out of hot cars and sunny windowsills.
If a lab asks for a replacement sample, that is routine, not a sign of anything wrong with you.
When blood or saliva can mislead
A few medical situations change whose DNA a sample contains, and they matter for clinical-grade results:
- Recent blood transfusion. Donor cells can contaminate a blood sample.
- Bone marrow or stem cell transplant. Blood and, often, saliva reflect the donor's immune cells, so a different tissue such as skin may be needed. Tell the ordering clinician.
- Blood cancers or clonal blood changes. Variants that arise in blood cells over a lifetime can show up in a blood test without being inherited. Clinicians sometimes confirm important findings in another tissue.
If any of these apply, talk to a clinician before choosing a consumer kit. See our guide to DTC versus clinical-grade testing.
How to choose
- Ancestry or curiosity: saliva or cheek swab is fine. Follow the instructions. Your ancestry percentages will depend far more on the company's reference data than on the sample type.
- Whole genome sequencing at home: saliva is common. Read the provider's sample-type notes and its policy on redraws, which belongs on your pre-purchase question list.
- A result that will guide treatment: ask your clinician for a clinical test, usually blood.
- Children and infants: cheek swabs are often used because they are painless, but testing children raises separate questions. See should parents test kids for adult-onset conditions.
What a lab does with leftover material is a separate question. We cover it in what happens to your saliva sample after you mail it in.
Bottom line
Collection technique, not sample type, causes most problems, and a replacement kit is a routine fix rather than a verdict on your DNA. Follow the kit instructions to the letter, choose blood when the result will guide medical decisions, and tell clinicians about transplants or transfusions before testing.
Common questions
Is a saliva DNA test as accurate as a blood test?
For common variants the agreement is high when the saliva sample passes quality control. In one study, more than 95% of variants found in blood were also found in paired saliva, with lower agreement for rare variants and copy-number changes.
Why was my saliva kit rejected?
Common reasons are too little saliva, contamination from food or drink, a missed stabilizer step, or low DNA yield. Labs usually send a replacement kit, and following the instructions closely improves the second attempt.
Can I eat or drink before giving a saliva sample?
Most kits say not to for a set period beforehand. The length of that period differs between providers, so check the exact wording in your kit.
Does sample type change my ancestry results?
Not in a meaningful way. Ancestry estimates depend on the company's reference panel and algorithm, not on whether the DNA came from saliva, blood or a cheek swab.
Should I choose blood for a medical test?
If the result will guide treatment, ask your clinician. Clinical labs typically use blood, and a clinician can account for transplants, transfusions or other factors that affect the sample.
Sources and verification
- Yao et al., BMC Medical Genomics (2020): quality of whole genome sequencing from blood versus saliva derived DNA in cardiac patients
- DNA Genotek (saliva kit maker): vendor material on sample types, read as a commercial source
Figures come from the 2020 study above, a single hospital cohort. The kit-failure causes and the transplant and transfusion points are general laboratory practice and vary by provider, so follow the instructions supplied with your own kit and ask your clinician about your situation.
Medical disclaimer. GenomeTesting.org publishes educational content about genetics and consumer genomic testing. Nothing here is medical advice, a diagnosis, or a substitute for care from a qualified clinician or genetic counsellor. Decisions about testing, treatment, or interpreting your own results should be made with a healthcare professional who can see your full history.