Hi Smith,

What does the density look like for the rotamers that are flagged for bad geometry? It could be possible that they are refining to low-probability conformations that are nevertheless justified by the data.

Also, how many residues does that 3% reflect? Is it 4 residues of 120 or 30 residues of 1000?

Shane Caldwell

McGill University

On Mon, Jul 13, 2015 at 2:20 AM, Smith Liu <[log in to unmask]> wrote:

Dear All£¬

By refinement from a 2.5 A crystal, the Ramachandran favoured is higher than 96%, Ramachandran outlier lower than 0.8%, but the rotamer outlier is about 3% which is relatively high. If I do the auto fit rotamer by coot, for most of the rotamer outliers auto fitted, nearby peptide bonds will be cleft, and if I do real space refine around the cleft bonds, although the cleft bonds can be connected, the rotamer outlier residues are still rotamer outliers. Will you please explain the possible reasons?

Smith