But I guess what I am doing here is questioning the idea that there is a robust relationship between certain genetic values and certain behavioural traits (those involved in mental illness).

I guess I think that social causal mechanisms are likely to be more robust than genetic causal mechanisms at the end of the day. The trouble is that we are more limited in finding out about social causal mechanisms than we are in finding out about genetic causal mechanisms because of practical and ethical constraints around manipulating aspects of society vs manipulating genes and / or neurology in rats.