Finding causal structures behind correlations is a key task of science. In biology, we have to do that for evolved, highly connected homeostatic feedback systems. How?
Moving from "pseudo steady state" assumptions to time series may be the only option.