no one would advocate for an early pregnancy test on the market that
can give a result 3 weeks earlier but comes with a 40% error rate. it
sounds like 23andMe did not adequately quantify their error rates or
prove their science according to FDA standards.
also, please don't misconstrue "genomic data" as automagically
factual. Sample prep and readout thresholds make a big difference,
amongst other sources of error. are they doing samples in triplicate?
are they doing 10 runs per gene fingerprint? are they just doing one?
what are their positive and negative controls and how often are they
run? I don't think anyone knows what their protocols are or how they
do QA/QC.
these are just growing pains of a new industry. it's not going to kill
23andMe, and I expect the consumer will benefit from this additional
scrutiny (though prices may increase).
jordan