A statistical approach reveals designs for the most robust stochastic gene oscillators