3 ms·
23AndMe is fine because it's conservative about what it reports, but genetics papers are generally very non-credible (as in, they haven't heard about the Credib
by astrange 3y ago
23AndMe is fine because it's conservative about what it reports, but genetics papers are generally very non-credible (as in, they haven't heard about the Credibility Revolution).
They constantly use tools that can only show correlation like GWAS and then just pretend it's causation. And yes, because of somatic mutations their favorite tool "twin studies" is not reliable.
(On the other hand, I'm not sure we want them doing intervention studies that could show causation, it seems kinda unethical.)
- rogerkirkness 3y agoTo be more specific, 23andMe shows your full exome which is where proteins are encoded in your DNA. Sort of like how having the entire codebase doesn't necessarily tell you what's going on in production with memory corruption, behavior on certain hardware, cache misses, etc. the analogy works well.