You can see that they assumed there is no correlation between genes and environment. I am wondering what will happen if they add an unknown correlation between the two in their model. Maybe that correlation is not estimatable from the twin experiment, due to the lack of degree of freedom. I haven't checked that yet.