Hi all:
I do not know if someone mentioned this problem here before. Recently I ran the AutoBuild to build the structure model, and it outputs a best model(R/Rfree 0.20/0.24). After that, I tried the program Phenix Refine using the .pdb and .mtz generated by AutoBuild. The R value increased to 0.23/0.30 by the end of this refine.
I wonder why this happened? Or if I have done something wrong?

(After AutoBuild, I modified the structure and added the substrate moleculars into the structure by hand.

I thought the increase of the R value was resulted from the inapposite position of the substrates.

But the R value still increased when I ran the Phenix Refine just by inputting the output of the AutoBuild without any manual modification.)

--