Hi,

For the input, the last line of PDB is as followng,
ATOM 21953 HH22 ARG H 295 53.862 115.032 84.116 1.00100.56 H

I would not recommend having H given that you are likely refining against >4A resolution map.

Clearly in each PDB of each model of the all states pdb produced, the b-factor is different, and they are different with the b-factors of the input pdb.

Thus will you please explain why there are 2 states pdbs which seems both are refined pdbs?

Not sure I understand this. I would say they are different because they went through refinement, so why it is unexpected to you?

In addition, is it normal that the single state refined pdb cannot be opended?

No. It must be a bug in your system.

Further, I have noticed that the phenix.real_space_refine run=ADP gave a resolution which is different from the mrc general resolution given by relion or some other software. Will you please explain base on what the phenix.real_space_refine run=adp gives the resolution?

If you do not provide resolution in input using keyword "resolution=.." it will try to estimate it given model and map. This may or may not be accurate, and the way it is estimated is different from any available software.

Hope this answers some of your questions!

Pavel