A particular class of  stochastic differential equations (SDEs) admit non-linearities in the sesne of McKean. This means that their coefficients do not only depend on time and on the position of the process solution but also on its marginal laws. Often they constitute probabilistic representation of conservative PDEs and can be used for applications to stochastic control. The possibility of approaching them with particle systems provides a Monte-Carlo type approximation of the mentioned conservative PDEs. In this talk we will illustrate how the method can be adapted to the case of a class of non-conservative PDEs. The talk is based on recent work with A. Le Cavil, J. Lieber and N. Oudjane.