From Charles Manski in the latest Journal of Economic Perspectives (pdf):
Someone reading empirical research relating human genetics to personal outcomes must be careful to distinguish two types of work. An old literature on heritability attempts to decompose cross-sectional variation in observed outcomes into unobservable genetic and environmental components. A new literature measures specifific genes and uses them as observed covariates when predicting outcomes. I will discuss these two types of work in terms of how they may inform social policy. I will argue that research on heritability is fundamentally uninformative for policy analysis, but make a cautious argument that research using genes as covariates is potentially informative.
From the same edition, Beauchamp and colleagues address the following question (pdf):
How, if at all, should economists use and combine molecular genetic and economic data? What challenges arise when analyzing genetically informative data?
I’ll post my thoughts on these articles when I have had a chance to digest.