Genotyping or sequencing: the difference in methods
The key points at a glance
Genotyping examines a predetermined list of individual positions in the genome. Sequencing reads a section letter by letter, without such a list. The difference is not one of thoroughness but of scope: a chip measures very accurately what is on its list and says nothing about anything else.
A chip reaches its conclusion through signal strengths and a statistical comparison. It is very accurate for common variants, but accuracy declines for very rare ones. All figures come from two peer-reviewed papers published in 2021, with the author and year shown directly where each figure appears.
First, you will read what a single position in the genome actually is. Next come the process on the chip, a comparison of both methods in Chapter 3, the accuracy figures in Chapter 5, and the question of why two saliva tests measure different positions. The limitations come at the end.
What to expect in this article
1. What an SNP is—and why it determines the question of which method is used
2. How a chip measures: a list of positions
3. Sequencing: reading instead of querying
4. Why the selection of positions determines everything
5. How accurately a chip measures—and where accuracy breaks down
6. What a chip cannot find
7. Why two saliva tests measure different positions
8. What this means for your own saliva test
9. Which method is used in a mybody®x saliva test
10. Which question requires which method
11. Limitations: what this article does not clarify
12. What matters in the method
Frequently asked questions
Sources
What an SNP is—and why it determines the question of which method is used
Your genome is a very long chain made up of four building blocks. At the vast majority of positions in this chain, all people have the same building block. But at some positions they do not: one person carries an A, while another carries a G. Exactly such a position is called an SNP, pronounced like the English word “snip” and short for single-nucleotide polymorphism—an individual position at which people differ by one letter.
Such positions are the raw material for every DNA analysis for nutrition, training, or skin care. It is not the entire genome that is examined, but the locations where differences between people exist at all.
Key message
An SNP is neither a trait nor a finding, but an address. Only the measurement method determines which of these addresses are actually examined.
How many such addresses there are
That number is the reason there are two methods at all. The public reference database dbSNP lists more than 100 million confirmed variants in the human genome (Lu and colleagues, 2021). No study in the world queries them all individually.
If no method tests everything, then every method tests a selection. This makes it interesting to examine how this selection comes about and who determines it.
Where the Sample Comes From Does Not Change the Method
A common misconception is that saliva is less accurate than blood. The Institute for Quality and Efficiency in Health Care describes several sample collection methods for genetic testing: usually cells from the blood, alternatively hair or cells from the inside of the mouth, for which a cotton swab is sufficient (IQWiG, as of 05.08.2026).
The genetic material is isolated from these cells using the same process and examined with highly precise methods. The sample therefore provides only the material. What is read from it is determined by the technology behind it—and that is precisely what the next chapter is about.
How a Chip Measures: A List of Positions
Genotyping uses what is known as an array, colloquially a chip: a small carrier on which a matching counterpart is placed for each position to be tested. When your DNA encounters its counterpart, a signal is generated. Where no counterpart is present, no signal is generated either—regardless of what is present at that location in your genetic material.
The briefest description of the method appears in a review in the British Medical Journal. It calls an SNP chip a DNA microarray that tests genetic variation at many hundreds of thousands of specific locations in the genome. The key phrase is specific locations: these locations are fixed before your sample reaches the laboratory.
Documented source
“SNP chips are DNA microarrays that test genetic variation at many hundreds of thousands of specific locations across the genome.”
Weedon MN, Jackson L, Harrison JW, Ruth KS, Tyrrell J, Hattersley AT, Wright CF
Use of SNP chips to detect rare pathogenic variants, BMJ 2021 (DOI 10.1136/bmj.n214)
The quotation is deliberately given in the original English. In German, it reads: SNP chips are DNA microarrays that test genetic variation at many hundreds of thousands of specific locations across the genome. The study comes from the field of medical diagnostics. This article adopts the measurement mechanism from it, not the medical examples.
How a Signal Becomes a Letter
A chip does not read letters. It measures signal intensities, and it is only from these signal intensities that one can infer which building block is present. According to Weedon and colleagues (2021), this is done through automated clustering: the signals from many people are considered simultaneously, and the groups that form are used to infer which DNA base is present in each person.
For groups to form, there must be enough people in each group. For a variant carried by many people, this is not a problem. For one carried by almost nobody, the group becomes so small that it can barely be distinguished from random noise.
Here is an analogy from the workplace: the chip works like a mailroom with fixed compartments. Each compartment has an address, and anything that arrives is reliably sorted into it. If mail arrives for an address without a compartment, it does not disappear into chaos—it simply never appears.
Sequencing: reading rather than querying
Sequencing takes the opposite approach. It does not ask about previously specified positions; instead, it reads through a section of the genetic material in sequence and reconstructs it from many short reads. No one needs to have anticipated what is found.
That is precisely why sequencing serves as the benchmark in research against which chip results are measured. Weedon et al. (2021) compared chip results with sequencing data from the same individuals to determine how often the chip was correct.
| Criterion | Genotyping | Sequencing |
|---|---|---|
| What is measured | signals at predetermined individual positions; the building block is inferred from the signals | the sequence of building blocks in a section, read out one by one |
| How many positions | for most products, 600,000 to 800,000; in practice, chip size is about one million (Lu et al., 2021) | all positions in the section that was read, regardless of whether they were previously known |
| What is and is not detected | what is on the list is detected; everything outside the list remains unmeasured | unexpected findings are also detected in the section that was read; outside the section, nothing is measured |
| What it is typically used for | broad studies of many people and analyses for consumers, which often run on arrays (NHGRI, as of June 14, 2023) | targeted questions about individual sections and as a benchmark in research (Weedon et al., 2021) |
The table contains no column for price or quality, and that is not an oversight. Which method is suitable depends on the question, not on a ranking.
Why the selection of positions determines everything
Someone decided which positions would be included on the chip. This selection is the substantive statement made by the method, and it cannot be expanded afterward.
The scale of this selection can be quantified. Most products measure 600,000 to 800,000 positions; in practice, chip size is capped at about one million (Lu et al., 2021). Compared with the more than 100 million confirmed variants in dbSNP, this amounts to less than one percent, according to the same study.
This claim is often repeated incorrectly, so let me state it explicitly: it is less than one percent of the known variants, not less than one percent of your genetic makeup. Those are two different claims, and only the first is supported by evidence.
Why the selection covers more than the number suggests
Positions in the genome are often inherited together. Anyone carrying a particular variant is also highly likely to carry certain neighboring variants. This regularity can be used to computationally infer what was not measured. This process is called imputation.
Imputation is why a selection of a few hundred thousand positions reaches beyond the edge of its own list. But it is not a measurement; it is a reasoned estimate. Lu and colleagues (2021) put its accuracy at under 91 percent even for the simplest case of common variants, depending on which positions the chip carries.
A measured position and an imputed position appear side by side in a report, with no visible indication of which is which.
Four terms that build on one another
What an SNP is
A single position in the genome at which people differ. dbSNP lists more than 100 million confirmed variants (Lu and colleagues, 2021).
What a chip does
It checks a predetermined list of such positions. The list is fixed before the sample arrives at the lab.
What sequencing does
It reads the section letter by letter, without a predetermined list. That is why it serves as the benchmark for chip results (Weedon and colleagues, 2021).
What the difference follows from
A chip cannot find what is not on the list. This is not a weakness of the measurement, but a property of its design.
How accurately a chip measures—and where that accuracy breaks down
Those who disparage chips understate their accuracy; those who oversell them understate the condition under which that accuracy applies. Both halves belong together.
The chip in numbers
600.000–800.000
positions measured by most products; chip size is practically capped at one million (Lu and colleagues, 2021)
over 99.9%
Estimated repeatability for today’s genotyping arrays: measuring the same sample twice produces the same result (Lu and colleagues, 2021)
over 99%
sensitivity, specificity, and positive and negative predictive values achieved for common variants (Weedon and colleagues, 2021)
Sources: Lu C, Greshake Tzovaras B, Gough J, Computational and Structural Biotechnology Journal, 2021 · Weedon MN and colleagues, BMJ, 2021
What these figures mean—and what they do not
The over-99-percent figures refer to 108,574 common variants; for these, chip results and sequencing agreed almost consistently across all four quality measures (Weedon and colleagues, 2021). For the vast majority of what a chip measures, the measurement is therefore highly reliable.
The repeatability of over 99.9 percent is often confused with this (Lu et al., 2021). It only means that measuring the same sample twice produces the same result twice. That is instrument stability, not an indication of whether the position measured was the right one for your question.
The condition under which accuracy declines
With very rare variants, the picture is reversed. Weedon et al. (2021) found that for variants with a frequency below 0.001 percent, only 16 percent of the 4,757 heterozygous genotypes reported by the chip could be confirmed in the sequencing data. Heterozygous here means that the chip shows two different building blocks at that position.
The reason is already stated in Chapter 2. Automatic clustering requires enough people in each group. For a variant carried by practically no one, the group is too small to separate it cleanly from the noise. The chip then provides an answer that is not statistically supported.
These two figures are often turned into one. The over 99 percent for common variants (Weedon et al., 2021) becomes the advertising claim, while the 16 percent for very rare variants (Weedon et al., 2021) becomes the takedown. Both simplifications leave out the condition under which each figure applies. Common variant, very accurate. Very rare variant, uncertain.
What a chip does not detect
The practically most important property of a chip is not an error rate but a blank spot. A position missing from the chip does not produce an incorrect result but no result at all, and in the report this spot looks exactly like an unremarkable one. There is nothing there.
A measurement error appears as a deviating value that can be checked. An unmeasured position does not appear at all.
Why this also applies to reports with no abnormalities
This leads to a rule for reading. An unremarkable section means that nothing unusual was found at the positions tested, not that there is nothing at untested positions. A report's evidentiary value ends at the edge of its list.
The National Human Genome Research Institute notes that reports from such tests, even with more comprehensive methods, contain only a predetermined set of variants (NHGRI, as of June 14, 2023).
Chapter at a glance
An SNP chip does not produce an incorrect result at positions it does not measure; it produces no result at all. Therefore, an unremarkable finding is always only a statement about the positions that were tested. The National Human Genome Research Institute notes that reports from such tests, even with more comprehensive methods, contain only a predetermined set of variants (as of June 14, 2023). Anyone who wants to interpret a result therefore needs the list, not just the finding.
Why two saliva tests measure different positions
If the list determines the method, then every list determines its own method. And lists are not standardized. The National Human Genome Research Institute puts it in one sentence: The number and location of the SNPs examined differ between providers (NHGRI, as of June 14, 2023).
Both matter, and the second is usually overlooked. Two chips can carry the same number of positions and still examine different positions. The number alone therefore says less than it appears to.
The same source also categorizes the methods. Consumer studies of health traits and carrier status often use SNP arrays; a few offerings use next-generation sequencing, but then for specialized questions (NHGRI, as of June 14, 2023). The standard in the consumer market is therefore the chip.
What this means for a comparison
Two reports that differ in one statement do not necessarily contradict each other. They may simply have examined different positions. Comparing two results is therefore only a comparison when the same positions are behind them. What follows when two reports actually diverge is explained in Two DNA Tests, Two Results.
The objection is obvious: Then every number on a package would be worthless. It does not go that far. It narrows down the order of magnitude, but it does not replace the question of which positions are involved.
What this means for your own saliva test
The entire technology can be reduced to one sentence.
A chip measures the positions it carries very precisely—and finds nothing at positions it does not carry.
A chip measures the positions it carries very precisely—and finds nothing at positions it does not carry. Neither half works alone. The first alone sounds like a promise; the second alone sounds like a warning. Together, they describe a tool.
For you as a buyer of a saliva test, two questions follow that you can ask before purchasing. How many positions are measured? And are the report's conclusions derived from measured positions or from positions supplemented through calculation? Both questions can be answered without anyone revealing trade secrets.
There is a third question, one that moves away from the method: What does the provider do with the positions? Two tests using the same chip can produce very different reports. Chapter 11 returns to this.
Which method is behind a mybody®x saliva test
What applies to chips in general also applies to the company's own tests. mybody®x (MYBODY Lab GmbH)'s DNA analyses use SNP genotyping at more than 700,000 positions from a saliva sample (as of 31 August 2026).
This places the method at the upper end of the range described for the market by Lu et al. (2021). It follows the same logic as any other chip, with the same gap outside the list.
DNA test from saliva
Longevity | ALL IN ONE DNA test
The Longevity test performs genotyping. It does not perform sequencing. More than 170 gene variants relating to nutrition, weight loss, fitness, and skin are evaluated and summarized in 74 reports. What it does not do: it does not read your genome in its entirety, it cannot find anything at positions outside the list, it does not provide a diagnosis, and it says nothing about diseases.
Laboratory analysis: 15–25 business days after receipt of the sample
Information from the product page, accessed on 31 August 2026
One figure from this chart deserves context that is rarely provided. The chip measures more than 700,000 positions, while more than 170 gene variants are evaluated. This is not a contradiction: not every measured position contributes to a finding supported by robust evidence.
How the number of measured positions is distributed among the individual reports has not been publicly documented. This information is therefore omitted here rather than estimated. It remains an open item for the editorial team.
Which question requires which method
For questions based on common variants, genotyping is the appropriate method. It measures these variants with accuracy rates above 99 percent (Weedon et al., 2021) and does so at many positions simultaneously.
For questions where a rare or previously unknown variant is decisive, a chip is the wrong tool. Not because it is inaccurate, but because the question lies outside its design. Such questions belong in clinical genetic diagnostics and therefore in the hands of a physician.
How to recognize this from your own question
There is a simple way to check. Are you asking how your body handles something that affects many people—caffeine, lactose, endurance exercise, nutrient requirements? Then you are dealing with common variants, and a chip is designed for that.
If, on the other hand, you are asking about a disease that occurs unusually frequently in your family, that is not a question for a consumer product. That is what genetic counseling is for; it uses different procedures, different requirements, and different conversations.
Limitations: what this article does not clarify
This article describes a measurement method. It does not describe how good the analysis based on that measurement is. Two providers using the same chip can produce very different reports, and the difference then lies not in the technology but in the interpretation.
The 16 percent from Chapter 5 (Weedon and colleagues, 2021) comes from a study concerned with rare disease-related variants. Its application to nutrition and lifestyle reports has not been established and is not claimed here. The only thing carried over is the mechanism explaining why clustering becomes weaker for very rare variants.
The imputation figure also has a limitation. The value of under 91 percent applies to the simplest case of common variants and depends on which positions the respective chip carries (Lu and colleagues, 2021). No figure for a specific report from a specific provider can be derived from it.
Three of the four sources for this article are in English and date from 2021 and 2023. Newer chip generations may accommodate different quantities. Where a figure is older than the current state of the art, its year appears directly beside it so that you can weigh it yourself.
And the open question from Chapter 9 remains open: How many of the measured positions feed into the individual reports is not publicly documented. An estimated figure would be the worse error.
What matters about the method
If you take away one action from this article, let it be this: Before comparing a saliva test, write down two questions and ask the provider. How many positions are measured? And which statements are based on measured positions, and which on computationally supplemented positions?
The reason is unremarkable. The number of reports, the length of the book, and the number of chapters describe how much was made from a measurement. They do not describe how much was measured. These are two different pieces of information, and only the second concerns the method.
At the beginning was the question of how genotyping and sequencing differ. The shortest answer is: in the list. One method has one; the other does not. Everything else in this article spells out this one difference.
Frequently asked questions
What is the difference between genotyping and sequencing?
Genotyping examines a predefined list of individual positions in the genome and infers from signal intensities which building block is present there. Sequencing reads a section letter by letter, without such a list, and therefore also detects unexpected findings in the section read. The scientific literature describes SNP chips as DNA microarrays that examine genetic variation at many hundreds of thousands of specific locations in the genome (Weedon and colleagues, BMJ 2021). The difference therefore lies in the scope, not in the care taken.
What is an SNP?
An SNP is a single position in the genome at which people differ in one building block: one person carries an A there, another a G. The public reference database dbSNP lists more than 100 million confirmed variants in the human genome (Lu et al., 2021). An SNP is not a trait or a finding, but initially just an address that a method may or may not target.
How many SNPs does a DNA test examine?
Most products measure 600,000 to 800,000 positions; in practice, chip size is capped at around one million (Lu et al., 2021). According to the same study, this corresponds to less than one percent of the more than 100 million confirmed variants in dbSNP—not one percent of the genome, which would be something different. The number and location of the positions tested vary between providers (NHGRI, as of 06/14/2023). The DNA analyses by mybody®x measure more than 700,000 positions (as of 08/31/2026).
How reliable is measurement on an SNP chip?
This depends on the frequency of the variant, and both figures belong together. For 108,574 common variants, sensitivity, specificity, and positive and negative predictive values were all above 99 percent. For variants with a frequency below 0.001 percent, however, only 16 percent of the 4,757 heterozygous genotypes reported by the chip could be confirmed in sequencing data (both Weedon et al., BMJ 2021). The repeatability of today’s arrays is estimated at over 99.9 percent (Lu et al., 2021).
How does a DNA test measure DNA from saliva at all?
Genetic tests generally use cells from blood, or alternatively hair or cells from the oral mucosa, for which a cotton swab is sufficient; the genetic material is isolated from the cells and examined using highly precise methods (IQWiG, as of 05/08/2026). The sample therefore only provides the material. Whether a list of positions is queried afterward or a section is read through is determined by the laboratory method, not by the type of sample.
Next step
The method first, then the decision
If you want to know how genotyping derives specific insights into nutrition, training, and skin from more than 170 genetic variations, you can find that on the product page. And if you are wondering whether a test taken once remains valid indefinitely, the answer is in the second article.
Longevity | ALL IN ONE DNA test Is one test enough for a lifetime?Read more
You might also be interested in
The level beneath this article: What the genetic material consists of before looking at measurement methods.
The follow-up question: which test method is suitable for which starting point.
Sources
- Weedon MN, Jackson L, Harrison JW, Ruth KS, Tyrrell J, Hattersley AT, Wright CF: Use of SNP chips to detect rare pathogenic variants: retrospective, population based diagnostic evaluation. BMJ 2021;372:n214, DOI 10.1136/bmj.n214 – bmj.com
- Institute for Quality and Efficiency in Health Care (IQWiG): What happens during a genetic test? (as of 08/05/2026) – gesundheitsinformation.de
- Lu C, Greshake Tzovaras B, Gough J: A survey of direct-to-consumer genotype data, and quality control tool (GenomePrep) for research. Computational and Structural Biotechnology Journal 2021, DOI 10.1016/j.csbj.2021.06.040 – pmc.ncbi.nlm.nih.gov
- National Human Genome Research Institute (NHGRI): Direct-to-Consumer Genetic Testing FAQ for Healthcare Professionals (as of 06/14/2023) – genome.gov
The verbatim English quotation defining an SNP chip, the information on automated clustering, the quality metrics of over 99 percent for 108,574 common variants, and the confirmation rate of 16 percent for 4,757 heterozygous genotypes of very rare variants are taken from [1]. The information on sample material and the procedure for a genetic examination is based on [2]. Chip size, repeatability, imputation accuracy, and the comparison with the over 100 million confirmed variants in dbSNP are taken from [3]. The fact that the number and location of the positions tested differ between providers, and that reports contain only a previously defined set of variants, is stated in [4]. Information on price, sample type, scope, and laboratory comes from the mybody®x product page, accessed on 08/31/2026; the processing time follows the central specification for DNA tests. All four sources were accessed and reviewed on 08/31/2026.
mybody®x Editorial & Specialist Team
Nutrigenetics Laboratory diagnostics Genetic analytics Nutritional science
This article was created by the mybody®x editorial and specialist team. The team combines nutrigenetics, laboratory diagnostics, and nutritional science. Those who contribute to it are listed on the authors page.
Published on 08/31/2026 · Last updated on 08/31/2026
The DNA analysis is intended for nutritional and lifestyle counseling. It is not a diagnostic procedure, does not predict diseases, and does not replace a medical examination or consultation. Genetic variants describe probabilities in population groups, not a predetermined outcome for individuals.
Share now:
Genotype and phenotype: the difference and what lies between them
Genotype and phenotype: the difference and what lies between them