> Because the variants have specific and large effects on alcohol, but do not effect other lifestyle factors such as diet, smoking, economic status or education

> Because the variants have specific and large effects on alcohol, but do not effect other lifestyle factors such as diet, smoking, economic status or education, they can be used by scientists to nail down causal effects of alcohol intake. Well, yes, but only in people with a similar genetic background. Using this as a kind of instrumental variable assumes an additive or linear model of genome environment interactions which is about as realistic as a flat Earth model. It may be that this study is being oversold.