Chip or full sequence: what your DNA test actually read
When you spit in a tube for 23andMe or AncestryDNA, your genome does not get "read". About 600,000 to 700,000 preselected positions get read, out of roughly 3.2 billion. That is around 0.02 percent of your genome, chosen in advance by the chip's designers.
Whether that is a problem depends entirely on the question you ask the data. This article is the difference between the two technologies, and the one big failure mode you should know about.
How a genotyping chip works
Consumer tests use a microarray, a chip dotted with hundreds of thousands of molecular probes. Each probe checks one known position in the genome: a spot where humans commonly differ, like rs762551 in the caffeine gene CYP1A2. The chip cannot discover anything new; it can only check the positions it was built to check.
For those positions, chips are excellent. For common variants, ones carried by at least a few percent of the population, chip accuracy is typically better than 99 percent, and those common variants are exactly what large population studies (and wellness reports like ours) are about.
Where chips quietly fail
The failure mode is rare variants, and it is worth being blunt because raw files are full of them.
In 2021, researchers checked chip calls against real sequencing in UK Biobank participants. For very rare variants, carried by fewer than about 1 in 100,000 people, more than 8 in 10 chip calls were wrong: the chip said the rare variant was there and sequencing said it was not. The math is unforgiving: when a variant is almost never truly present, even a tiny error rate means most positive calls are false.
The practical consequence: if you spot a rare, scary-looking mutation in your raw file, in BRCA1 or anywhere else, the odds are good it is a chip artifact. It needs confirmation by clinical-grade sequencing before anyone should act on it, and reputable services will tell you exactly that.
What whole-genome sequencing does differently
Whole-genome sequencing reads essentially all 3.2 billion positions, typically 30 times over for accuracy. It can find variants nobody put on a chip, which is why it is the right tool for diagnosing rare diseases, and why clinics use it. The trade-offs are cost, heavier data, and, for most wellness questions, no added benefit: the variants with solid evidence for traits like caffeine response, chronotype, or body weight are common, and the chip already reads them well.
The honest summary
- Chip: reads a curated 0.02 percent, superb at common variants, untrustworthy on rare ones.
- Whole genome: reads everything, the tool for rare and clinical questions, more than you need for common-variant wellness traits.
Helisoma is built deliberately on the first case. Your report covers common, replicated variation, the kind your chip measures reliably and large studies actually support, and it stays away from rare clinical calls that chips get wrong. Upload the raw file you already have, and get the analysis that file is genuinely good for.
Sources
- Weedon MN et al. Use of SNP chips to detect rare pathogenic variants: diagnostic evaluation in UK Biobank. PubMed 33589468