What can genetics tell us about environment?


In our G=E paper we tried to highlight what
we think is an underappreciated fact about genome wide association studies; that not
only do they tell us about underlying biology, but they can also tell us about modifiable
or environmental causes of disease. The best illustration of this I think is that
when you run a genome wide association study of lung cancer for example, you identify a
signal on chromosome 15 and that signal seems to be driving not lung cancer directly but
our risk of smoking heavily for a smoker. In other words, what that genetic influence
is acting on more proximally is whether or not we smoke and because smoking is causually
linked to lung cancer, when we run a genome wide association study of lung cancer we see
that signal for a modifiable risk factor for that disease. I think that insight is important because
it gives us a perspective on genome wide association studies that we might otherwise miss. So to illustrate that, a recent study of schizophrenia
also indicated a signal on chromosome 15 in the same signal seems to be driving smoking
behaviour. Now we know that smoking and schizophrenia
are highly related and the assumption has always been that for the most part that that
relationship is due to people with schizophrenia smoking to self-medicate they symptoms or
the side effects of their medication; but one possibility that this different perspective
gives us is that smoking may be in fact a causal risk factor for schizophrenia. And there have been some other studies since,
a meta-analysis out of Kings College London and a Mendelian randomization study
out of Copenhagen that have added further support to this possibility. And another analysis from some Swedish data
by Ken Kendler and colleagues, all pointing to the possibility that smoking may be a causal
risk factor for schizophrenia and so this perspective on how we interpret genome wide
association studies might give us insights not just into the underlying biology, but
into other risk factors that we might be able to act on much more rapidly.

Add a Comment

Your email address will not be published. Required fields are marked *