The main plot shows a histogram of the differences between the actual and predicted chemical shift values, for the type of atom selected by the cursor in the bottom plot. For CA and CB, values from all residues are shown; for other atoms, only values from residues with helix and sheet secondary structures are shown.
The average and tolerance for CA and CB include values from all residues; the average and tolerance for all of the other atoms include only values from residues belonging to helices or beta strands.
All GLY HA2 and HA3 assignments are excluded from the plots and calculations. All shifts from residues with large backbone B-factors (1.6 larger than the average) in the corresponding X-ray structure are excluded from the plots and calculations.
The N predicted shifts of SER, THR and GLY are excluded from all calculations.
All chemical shift referencing correction calculations exclude large outliers, i.e.
For all atoms a chemical shift referencing correction is needed (advised) if the average deviation is larger than
3*rmsd(atom)/sqrt(N(atom))
i.e., only if the average deviation exceeds by more than a factor of three the expected random variation in this average. The width of the brown area represents this maximum expected average deviation and the suggested correction corresponds ONLY to cases when the origin of the chemical shift difference axis lies outside the brown area. Only in these cases is there a need for a referencing correction in the amount (in ppm) indicated by the position of the thick orange line (the average deviation).
The chemical shift referencing correction method uses chemical shift values simulated by the DC program included in the NMRPipe package ( http://spin.niddk.nih.gov/NMRPipe/dynamo/).