The calculation of molecular conductance in the density functional theory is inaccurate, and the formalism needs to be extended. The GW approximation to the many-body perturbation theorey is presented, together with the formalism of inclusion of coupling to macroscopic reservoirs.