The False Positive Epidemic in Raw DNA Data
The terms of service explicitly state raw data is "unvalidated for medical use." When you ignore this warning and run it through third-party parsers, the results can be terrifying—and frequently wrong.
It is a common scenario in genetic counseling clinics today: a panicked patient arrives holding a printout from a website like Promethease or Genetic Genie, claiming they have a rare, highly pathogenic mutation for breast cancer or a severe metabolic disorder.
The clinician orders a medical-grade genetic test (often using Sanger sequencing) to confirm. In up to 40% of these cases, the dangerous mutation simply isn't there. It was a false positive.
Why Do Microarrays Fail at Rare Diseases?
Consumer DNA tests like AncestryDNA and 23andMe use a technology called microarray genotyping. They do not sequence your entire genome letter by letter. Instead, they use a chip with chemical probes designed to bind to specific, known variants (SNPs) scattered across your genome.
- Optimized for the Common: These chips are brilliant at detecting common variants used for ancestry tracing or common traits (like eye color).
- Terrible at the Rare: The probes can be "noisy." Sometimes a probe binds weakly to the wrong piece of DNA. If a disease-causing variant is extremely rare in the general population, the mathematical probability that a positive signal on a microarray is a false positive is overwhelmingly high.
The 2018 Ambry Genetics Study
A landmark study published in Genetics in Medicine evaluated 49 patient cases where individuals sought clinical confirmation of a pathogenic variant discovered in their DTC raw data. The researchers found that 40% of the variants reported in the raw data were false positives. The false positives were particularly concentrated in genes related to hereditary cancer (like BRCA1/2, CHEK2, and ATM).
The Responsibility of Third-Party Tools
Third-party health analysis sites operate in a regulatory gray area. They argue they are merely "literature retrieval" services—they take your raw data file and cross-reference every SNP against a database like SNPedia, spitting out the associated medical literature.
Because they assume every data point in your raw file is 100% accurate, they will surface catastrophic health warnings based on microarray errors. They lack the clinical context to recognize when an artifact is biologically implausible.
What To Do Instead
If you have a personal or family history of a genetic condition, skip the consumer spit kit entirely. Speak to a board-certified genetic counselor who can order a clinical-grade panel tailored to your specific risk profile. If you have already used a third-party tool and received alarming results, do not make medical decisions based on it. Seek clinical confirmation.