Hi Wei,
Thank you guys so much for the suggestions! I don't know why I still have rotamer outliers after real-space refinement. I am wondering what might go wrong and what I should check in this case?
which Phenix version you use? Anyway, sounds like you don't have rotamer outliers as statistics below indicates.
At the same time, I fixed the Ramanchandran outliers and rotamer outliers in Coot, and did another refinement with default setting + NCS restraints + secondary structure restraints + optimize X-ray/ stereochemistry + optimize X-ray/ ADP. Are these refinement settings look reasonable to you? Below are the results: R-work: 0.2542, R-free: 0.2857, Angles: 1, Bonds: 0.003. Ramachandran outliers: 0.4% (Goal: <0.2%), Ramachandran favored: 92.6% (Goal: >98%). Rotamer outliers: 0 % (Goal: 1%), c-beta outliers: 0 (Goal: 0) clashscore: 6.92, overall score: 1.85.
Looks ok to me.
But, when I opened the results in coot, do Ramanchandran plot, and I couldn't see any Ramanchandran outliers.
This is because Coot and Phenix use slightly different Ramachandran plot libraries. I guess Paul Emsley or/and Jeff can offer more comments on this. Pavel