cited research aside this is asinine. let's pretend we could regress "did smile" on "genetics" and "culture" and inspected the coefficients of the covariates (i.e. investigate treatment effect). we would find that the effect of culture was much much much stronger than genetics.