The process of protein structure determination based on NMR data can be divided in several stages:

1. Data Acquisition
2. Data Processing
3. Resonance assignments
4. Determination of structural constraints from experimental data such as NOEs, J, dipolar couplings an others. NOEs and dipolar couplings are generally transformed into distance constraints and scalar J data converted to dihedral angle constraints
5. Generation of structures using Distance Geometry (DG), simulating annealing(SA) or other techniques.
6. Refinement of these structures using a combination of restrained energy minimization and molecular dynamics (MD) simulations (02JB317-22 ).

Different approaches have been used to improve the precision and accuracy of a structure determined by NMR:

Distance restraints obtained from NOE data are the usual way to work. However, there is a problem when NOE data is limited as in the case of deuterated proteins.

Direct refinement against chemical shifts ( 98PNAS5891 and 95JMRB92-106 ), coupling constants ( 94JMRB99-104 ), and a conformational database potential ( 97JMR171-125 ).

Inclusion of residual dipolar couplings restraints for structure determination and/or refinement of proteins and protein-ligand complexes even when minimal NOE data sets are employed

Structure refinement using 13CO CSA ( 01JACS11065 )

Protein Structure elucidation using the assignment-independent CLOUDS approach ( 02PNAS6707 and 05METH261 )

Structure Validation using the radius of gyration ( 99JACS2337 and 01JACS3834 )

Use of distance-dependent line broadening effects originated from site-directed spin labeling.

Joint refinement between NMR and X-ray data ( 01JB235-21 )

Improving the quality of protein structure derived by NMR ( 02JB281-22 )

Identification of protein NOESY spectra via unassigned spin systems ( 02JB203-24 and 04JB1-28 )

General Reading: 96ENC3801 , B03ZER39 and 03PROG105