Dear Meytal,

Perhaps the problem can be solved by increasing the restraint weight in Refmac, perhaps your model needs more substantial rebuilding. Look at hydrogen bonding, you may need to flip a few peptides. Also a wrongly fitted side-chain may distort the backbone enough to cause a Ramachandran outlier.

Shameless plug: Try the pdb-redo.eu server to see if that can find good Refmac settings. The homology restraints in pdb-redo may be triggered for your model (depends on your data quality) as well. These too affect the Ramachandran plot.

Cheers,

Robbie

Sent from my Windows 10 phone