New blogpost: CORRELATION, PREDICTION AND EXPLANATION IN THE GENOME. A (Part II) discussion of @RobertPlomin and Sophie von Stumm on GPS for intelligence in @NatureRevGenet. @kph3k @charlesmurray @lucasjmatthews. https://goo.gl/mYaUnr (Previous post omitted the link)
-
If the R squared is large enough, you can predict outcomes and do genomic engineering (esp. embryo selection) without knowing anything about mechanisms. That's the real revolution. Eugenics is back. It's good again.
Actually, predictive validity depends on the R, not R2. Important because R is a fast growing nonlinear transformation of R2.
