If you find that the normal load on those now-frictionless supports is low, then your modified model is probably OK.
You may want to re-run the analyses that worked on the original job (under a different name) and compare results with the new model. If the common load cases yield similar results, then, again, you're probably OK.
I would caution against changing the default friction. Years ago we had a very low "Friction stiffness" and that did not generally agree with users' expectations. If you do reduce friction stiffness, you should check to see how much the "non-sliding" supports actually slide. These "non-sliding" supports will carry less than mu*N friction load but will moving anyway (due to that lower friction stiffness).
Another thing you might try to see what's going on is to simply reduce the friction multiplier in the Load Case Options. This will allow you to adjust the friction coeficient on a load case-by-load case basis.
_________________________
Dave Diehl