Abstract. 
We consider the asymptotic behavior of stochastic differential equations.
Instead of the traditional large-deviation approach, we study the
Fokker-Planck equation on the level sets of reference functions. New
estimations of invariant measures for stochastic differential equations
will be presented. We also study applications in biological networks. Some
systematic features of complex biological networks can be quantified by
activating the network with external noise.