The numerical solution of the Chemical Master Equation (CME) governing gene regulatory networks and cell signaling processes remains a challenging task due to its complexity, exponentially growing with the number of species involved. When considering separated representations of the probability distribution function within the Proper Generalized Decomposition-PGD-frame-work the complexity of the CME grows only linearly with the number of state space dimensions. In order to speed up calculations moment-based descriptions are usually preferred, however these descriptions involve the necessity of using closure relations whose impact on the calculated solution is most of time unpredictable. In this work we propose an empirical closure, fitted from the solution of the chemical master equation, the last solved within the PGD framework.