I used first principle calculations to study the thin film model. An obvious crossing was shown to appear at the gamma point when the edge state was calculated, which was originally predicted to be the evidence of quantum spin hall phases. The edge state occur at the Gamma point but it is not the location of of VBM. Therefore when I use the recursive green function method to calculate the edge state, the crossing appear inside the bulk band region. If we use the non-Abelian Berry connection method to calculate the $Z_2$, is it possible to get a $Z_2$ = 0 result in this situation?