4 ms·

"risk exposing people’s genetic health information..." Genealogy DNA tests only use 0.07% of a person's DNA (that's 0.0007 or seven ten-thousandths). Plus it's

"risk exposing people’s genetic health information..." Genealogy DNA tests only use 0.07% of a person's DNA (that's 0.0007 or seven ten-thousandths). Plus it's in the non-coding or "junk" zone of the chromosomes. Not much opportunity to figure out people's health issues from that little DNA.

sct202 7y ago
Plus, there are issues with false positives with the consumer grade testing. https://www.nature.com/articles/gim201838 https://www.nature.com/articles/gim201838
- eddyg 7y ago
  Companies like 23andMe regularly improve their technology and offer upgrades to get improved results. On this page [0], you can scroll down to "23andMe Health Reports by Chip Version" and select "See All" to see the changes. [0] https://customercare.23andme.com/hc/en-us/articles/218392668-Upgrading-to-23andMe-s-Newest-Platform https://customercare.23andme.com/hc/en-us/articles/218392668... (I'm not saying that their aren't false positives, just that improvements are being made as technology improves.)
- mnw21cam 7y ago
  My colleague is just bringing out a paper that shows that for variants that are rare in the population, the techniques used by 23&me and similar produce results that are almost exclusively false positives. https://www.biorxiv.org/content/10.1101/696799v1 https://www.biorxiv.org/content/10.1101/696799v1
- searine 7y ago
  The thing is, there are so many sites being sampled that any one false positive is buried under a mountain of true positives. False positives only matter when you are trying to be definitive about disease diagnosis. That is why clinicians follow up with more sensitive DNA tests after something like 23 and me detects a deleterious variant. For something like ancestry, the number of true positives is so ridiculously large that a few false positives don't matter.
kingkawn 7y ago
The idea of “junk” dna has been invalid for about a decade
searine 7y ago
>. Plus it's in the non-coding or "junk" zone of the chromosomes. The site themselves aren't important but they are physically linked to important sites. DNA is a chain, if you which section you are in and roughly what it looks like you can infer someones disease state.

sct202 7y ago

Plus, there are issues with false positives with the consumer grade testing. https://www.nature.com/articles/gim201838 https://www.nature.com/articles/gim201838

eddyg 7y ago
Companies like 23andMe regularly improve their technology and offer upgrades to get improved results. On this page [0], you can scroll down to "23andMe Health Reports by Chip Version" and select "See All" to see the changes. [0] https://customercare.23andme.com/hc/en-us/articles/218392668-Upgrading-to-23andMe-s-Newest-Platform https://customercare.23andme.com/hc/en-us/articles/218392668... (I'm not saying that their aren't false positives, just that improvements are being made as technology improves.)
mnw21cam 7y ago
My colleague is just bringing out a paper that shows that for variants that are rare in the population, the techniques used by 23&me and similar produce results that are almost exclusively false positives. https://www.biorxiv.org/content/10.1101/696799v1 https://www.biorxiv.org/content/10.1101/696799v1
searine 7y ago
The thing is, there are so many sites being sampled that any one false positive is buried under a mountain of true positives. False positives only matter when you are trying to be definitive about disease diagnosis. That is why clinicians follow up with more sensitive DNA tests after something like 23 and me detects a deleterious variant. For something like ancestry, the number of true positives is so ridiculously large that a few false positives don't matter.

kingkawn 7y ago

The idea of “junk” dna has been invalid for about a decade

searine 7y ago

>. Plus it's in the non-coding or "junk" zone of the chromosomes. The site themselves aren't important but they are physically linked to important sites. DNA is a chain, if you which section you are in and roughly what it looks like you can infer someones disease state.