Based on Your Search or Genetic Results?

Yahoo Japan is to market a genetic test.

Full-text access for registered users only. Existing users login here.
New to GenomeWeb? Register here quickly for free access.

Well that isn't different

Well that isn't different than the US where Big Sister of 23 and Me is in bed with Big Brother of Google. And yes I believe there are deliberate plans in place to crosslink and dossier all databases of information on all accessible individuals, to maximize the commercial value of everything they can "predict" (with ofensively intrusive, privacy invading, low predictive value algorithms) and sell to other sneaks. Brave new world eh.

Beg to differ. The AUC of

Beg to differ.

The AUC of predicting breast cancer using genetic information is about 0.65, while the AUC of using cholesterol level to predict coronary heart disease (CHD) is slightly better, 0.67? As I won't call CHD prediction using cholesterol "low predictive value", I won't call genetic testing "low predictive value", either.

Also, think about the case of BRCA1. The MAF of BRCA1 mutations in Jews are no more than 0.3%. Pretty low, huh? For the general population, using BRCA1 to predict breast cancer seems to be of "low predictive value" because the AUC will be close to 50.0% (note there is a decimal point). Then what's the use of BRCA1 testing? Myriad charges $3999/sample, and in some cases, insurance companies pays. I think that says something. For breast cancer, BRCA1 testing is low in sensitivity, but high in specificity. Predictive value is low -- AUC would say it's "nonsense". Yes, but for people with BRCA1 mutations, the predictive value is quite high, isn't it?