Resonances and Assignment

An overview of the CCPN data model assignment system, with teh core 'Resonance' concept.

What does 'resonance' mean in the Data Model?

The Data Model separates the assignment of NMR information and the assignment to atoms, so that it is possible to group NMR information arising from the same atom(s) without knowing which atom(s) are involved. The Data Model handles this via the Resonance object (see figure below). A Data Model Resonance transcends the traditional meaning attributed to the word in NMR: one Resonance object can have many different shift values belonging to different experiments. 

In other words, the Resonance object links all NMR information, and makes it possible to directly connect chemical shifts, peak assignments, constraints, ... . In this way the assignment of the NMR data and the assignment to atoms are separated. This also means that when the atom assignment for a Resonance is changed (e.g. a different stereospecific assignment), all NMR information also automatically 'knows' about this new atom assignment state.

An example of Resonances at work...

Consider the following example: you have identified two separate signals in NMR Spectrum 1, and assigned the first one to Resonance 'A' (blue), the second one to Resonance 'B' (red). In spectrum 1, Resonance 'A' has chemical shift value 7.89 ppm, Resonance 'B' is at 3.22 ppm. At this stage you know nothing about which atoms these Resonances correspond to.

You then peak pick Spectrum 1. You know that for peaks 322 and 323 the first dimension (dim. 1) corresponds to the signal at 3.22 ppm (in Spectrum 1), so you connect the first dimension for these peaks to Resonance 'B'. Similarly, for peak 377 you know that the third dimension (dim. 3) corresponds to the signal at 7.89 ppm, so you connect the third dimension for this peak to Resonance 'A'.

At this stage, you know the following:

• Resonance 'A':

• Chemical shift value of 7.89 ppm in Spectrum 1

• Assigned to the third dimension of peak 377 in Peak list 1

• Resonance 'B':Chemical shift value of 3.22 ppm in Spectrum 1

• Assigned to the first dimension of peaks 322 and 323 in Peak list 1

You now record Spectrum 2. Because you recorded this spectrum at, for example, a different temperature or pH, it has changed compared to Spectrum 1. By comparing the two, you see that the signal at 3.22 ppm is now at 3.26 ppm, and the signal at 7.89 ppm has shifted to 8.03 ppm. The signal, of course, still comes from the same source, so you connect the chemical shift value of 8.03 ppm to Resonance 'A', and the chemical shift value of 3.26 ppm to Resonance 'B'.

At this stage, you know the following:

• • Resonance 'A':

• • • • Chemical shift value of 7.89 ppm in Spectrum 1

      • Chemical shift value of 8.03 ppm in Spectrum 2

      • Assigned to the third dimension of peak 377 in Peak list 1

• • Resonance 'B':

• • • • Chemical shift value of 3.22 ppm in Spectrum 1

      • Chemical shift value of 3.26 ppm in Spectrum 2

      • Assigned to the first dimension of peaks 322 and 323 in Peak list 1

You then start a sequential assignment based on the information you have, and find out that Resonance 'A' corresponds to the HN atom of residue 11 in chain A. You connect Resonance 'A' to this Atom in the Data Model, and this way assign it unambiguously. All the NMR information you previously linked to Resonance 'A' is now connected to the correct atom. For Resonance 'B', you do not have a definite assignment, but you think it is the HB atom of residue 15 in chain A, so start using 'A.15.HB?', a plain text string, as the name of Resonance 'B'. This is not the same as connecting it to the HB atom of residue 15 in chain A, as 'A.15.HB?' is only a string, and though it might carry useful information for the user, it does not unambiguously describe the Atom assignment of Resonance 'B'. You might later decide that Resonance 'B' actually corresponds to atoms HD1 and HD2 of residue 25 - once you connect Resonance 'B' to these Atoms, this is all you need to know, and the name of the Resonance ('A.15.HB?') is meaningless.

The situation is then the following:

• • Resonance 'A', corresponding to atom HN in residue 11 of chain A::

• • • • Chemical shift value of 7.89 ppm in Spectrum 1

      • Chemical shift value of 8.03 ppm in Spectrum 2

      • Assigned to the third dimension of peak 377 in Peak list 1

• • Resonance 'A.15.HB?':

• • • • Chemical shift value of 3.22 ppm in Spectrum 1

      • Chemical shift value of 3.26 ppm in Spectrum 2

      • Assigned to the first dimension of peaks 322 and 323 in Peak list 1

If you would then make a constraint list based on the information coming from your spectra, you also link the constraint item members to Resonances. In this case, however, the Atom assignment situation is stored separately at the time of the creation of the constraint list, so that a constraint list never changes. If you change your mind about an assignment, you have to create a new constraint list.

How is it used in the FormatConverter?

In the FormatConverter, all NMR related information that is imported is first linked to resonances. The original chain/residue/atom names from the external format are used as the temporary 'key' for each resonance.

On the other hand, sequence information is read in from an external format to create molecules, a molecular system, chains, residues and atoms in the Data Model

At this stage, you have on one hand the NMR information, linked to Resonances, and on the other hand information about all the Atoms in the molecular system. There are as yet no links between the Resonances and the Atoms.

The linkResonances script (also part of the FormatConverter suite) is then used to link Resonances to Atoms. Once this is done, you can export the data in any external format, as the Resonance-Atom link is now unambiguously described.