Figure 1. NMR experiments yield a wealth of indirect structural information from which the protein three‐dimensional structure can only be revealed by consistent interpretation of the separate and distinct sets of NMR spectra recorded. Protein NMR structure determination thus entails building atomic‐resolution models that must simultaneously fulfill all experimentally determined conformational restraints. The principal source for the collection of conformational restraints is derived from the NOE that measures internuclear distances between hydrogen atoms in close proximity. The figure shows the standard, stepwise‐applied NMR structure determination protocol that can be divided into two principal data analysis components: (a) a first set of NMR spectra used for sequence‐specific resonance assignment and (b) a second set of NOESY spectra used to probe for spatial proximity of hydrogen atoms. A protein NMR structure is finally represented as a bundle of conformers that are all equally well satisfying the dense network of NOE‐derived distance restraints.

Figure 2. Flowchart of the entire UNIO protocol comprising expert systems for all stages of the NMR structure determination process.

Figure 3. Illustration of the UNIO output criteria for judging the correctness of the resulting NMR structure bundle. For reliable automated NOESY analysis, the initial three‐dimensional structure obtained at the outset of a structure determination (cycle 1) should be reasonably compatible with the experimental input data and show a defined fold of the protein. Structural changes between the first and subsequent ATNOS/CANDID cycles should only occur within the conformation space determined by the initial bundle of conformers obtained after ATNOS/CANDID cycle 1.