Main Panel¶
Peak List: Selects which peak list is used in the generation of distance restraints
Clone: Clone popup window
Help: Show popup help document
Close: Close popup
Create Structural Distance Restraints From NMR Peaks
This system is used to make resonance to resonance distance restraints based on peak intensity information from through-space NMR experiments, such as NOESY. Three mechanisms are available with regard to the assignment of resonance pairs to individual restraints, but for all three methods the distance information is extracted from peak intensities in the same way. The use of peak intensities and the three restraint generation methods are described below.
Restraint Distance Estimation
The options available in the “Restraint Distance Params section” if the “Settings” tab allow the user to specify how through-space peak intensities relate to the distance bounds of any any generated distance restraints. The default method is to calculate a target distance as peak height raised to the power of -1/6 multiplied by some scaling factor. However the user may also use peak volumes by changing the “Intensity Type” and other power relations via the “Distance function” pulldown menu.
The peak intensity scaling factor is defined such that the reference intensity (defaults to the peak list’s average volume) exactly corresponds to the reference distance (default 3.2 Angstroms). The user may change either of these values to calibrate in a different way. For example by measuring the peak intensities of known atomic distances such as those in an aromatic ring. The default values are appropriate for most protein structure calculations based on NOESY experiments, and in any case structure calculation programs like ARIA and CYANA will normally re-calibrate peak intensity scaling during the structure generation process. Hence, the initial calibration is not too critical; covalent geometry will provide better global calibration.
The upper and lower bounds of the distance restraints are calculated as fractional differences from the calculated target distance. For example if a target distance is 3.2 Angstroms an upper fractional error of 0.20 gives an upper limit of 3.84 (20% more) and 0.2 lower fractional error gives a lower limit of 2.56 (20% less). The upper and lower bounds are also confined the upper and lower distance limits. These are absolute minimum and maximum values for the bounds (1.72 & 8.00 Angstroms respectively by default).
If none of the fractional power distance functions are appropriate for the NMR data then the user may select the “Distance Bins” option. Here the user defines peak intensity classes, each of which relates to a target distance and distance bounds. Each class has a “Min. Peak Intensity” value, which is a multiplying factor relative to the reference intensity (by default the peak list mean). When Analysis determines which distance class to use for a peak it first divides its intensity value by the reference intensity to get a relative value, then the distance classes are searched for the class with the largest “Min. Peak Intensity” that is below the peak’s relative intensity. For example with two classes defined at 0.0 and 1.0 minimum intensities, a peak with a relative intensity value of 0.5 would fall in the first class, and a peak with 1.2 would fall in the second.
Restraints From Fully Assigned Peaks
The simplest way of generating distance restraints from through-space NMR experiments is to use the spectrum peaks that have been fully assigned. If a peak is caused by correlation between resonances and the atomic identities of those causal resonances is known (or at least suspected) a distance restraint can be created from the peak information. In CCPN terms the restraint will be between the resonances that are assigned to the through-space correlated peak dimensions, but only if those resonances are assigned to particular atoms (in a molecular system). Restraints are made for assigned peaks by selecting an appropriate peak list in the pulldown menu, at the very top of the popup, above the tabs, and pressing the [Make Assigned Restraints] at the bottom. Naturally, the user should make sure that parameters are specified for distance estimation and residue ranges and that the correct restraint set is selected in the “Settings” tab, to say where the generated restraint list should be placed.
Restraints will be generated for peaks that carry multiple, ambiguous assignments. This happens where the observed peak in a spectrum covers multiple resonance intersections that are unresolvable due to the relative scale of linewidths compared to the similarity of chemical shifts. In such circumstances all of the resonances assigned to the peaks through-space correlated dimensions will appear in the restraint. By default, if a peak is assigned to resonances A & B in one dimension and C & D on the other then an ambiguous restraint will be generated with items for all four resonance pairs; A-C, A-D, B-C and B-D, and a restraint would be satisfied by any of these pairs being within distance bounds. However, if the user knows that only certain pairings are possible on an ambiguously assigned peak then the assignments may be grouped via the Assignment Panel. Referring to the previous example if resonance A & C are in group 1 and B & D are in group 2 then the generated restraint will only have two items; A-C and B-D.
Restraints From Matching Chemical Shifts
There is a second common way to generate distance restraints, which is to match the chemical shifts of resonances to unassigned peak positions, thus generating potentially highly ambiguous distance restraints. Such restraints would typically be filtered to select only the correct contributing resonance pairs, by iterative structure generation and violation analysis in a program like ARIA or CYANA. Often the user can leave the matching of chemical shifts to the structure generation program, by passing the program peak lists rather than restraint lists. However, it is also possible to make such restraints in CCPN. The [Make Shift Match Restraints] can be used to generate restraints for the unassigned peaks in the selected peak list, however the user should be aware of all of the parameters that will be applied, including the “Shift Match Tolerances”, “Residue Ranges” and “Chem Shift Ranges”.
The “Shift Match Tolerances” tab is important because it specifies how the matching is done between (unassigned) peak positions and resonances with known chemical shifts. In essence, every unassigned peak dimension is compared to the chemical shifts contained within the shift list that is associated with the NMR experiment from which the peak derive. The shift list that relates to an experiment, and hence to a peak list, may be adjusted in the main Experiments table. Resonances whose shifts are close to the peak position may be the cause of that peak and so can be used to make a restraint. Often several resonances will match in each peak dimension, resulting in a restraint with many alternative, ambiguous resonance pairings (restraint “items”). The incorrect possibilities may be removed by automated structure generation programs, which generate guide structures based on the unambiguous restraints. The ”... PPM tolerance” values are used to control the maximum allowed difference from a peak location to a chemical shift for a resonance to be considered as an assignment possibility; the threshold width of the shift matching. In the absence of linewidth information only the “Min/default PPM tolerance” is used. If a peak has a recorded linewidth then its value may be used to scale the shift-match threshold, i.e. wider peaks match a wider PPM range. In such cases the shift-match threshold is the peak’s linewidth multiplied by the “line width multiplier” value, but bounded by the minimum and maximum PPM tolerance values.
The [Test Shift Match] function can be used to trail the peak to resonance matching procedure without generating any distance restraints. This is useful to test how many peaks would fail to be matched, and the resonance why.
When shift-match restraints are made the resonance matches can be filtered by using a guide structure (e.g. homologue or rough structure) with a corresponding distance cutoff, so that resonance pairs that are far apart in the guide will not be places in the restraints. Also, shift matching obeys any isotope labelling scheme setting; any resonance pair that has a spin-active isotope incorporation below the “Minimum Isotope Fraction” is deemed not to be visible and will not be present in a restraint.
Network Anchoring
In the case of the shift-matching method potentially ambiguous distance restraints are generated by simply matching peak positions to close chemical shifts. In the case of network anchoring method, chemical shifts are also matched to peaks, but the ambiguous possibilities are refined by selecting only NOE assignments from amongst the possibilities that are supported by other known close resonances or covalent structure. Say, for example, that a peak could have arisen from a number of resonance pairs with matching shifts. Two resonances A & B are more likely to be a correct assignment for the peak if we know that they are close to (or bound to) the same intermediary resonance, C and therefore must be close to each other.
The network anchoring procedure is directed using the “Network Anchoring” tab and operates over several peak lists at the same time. These peak lists are selected in the “Peak List Selection” table by double clicking in the “Use?” column to indicate “Yes”. Any assigned peaks in these peak lists, together with the covalent connectivity of the atoms, will form the initial through-space network. The restraint generation occurs after the iterative network refinement process converges (usually in fewer than five cycles). The parameters for the refinement are specified above the main table. It should be noted that only a shift list and set of shift match tolerances is currently used.
The “Min net. score” is used during refinement to dictate how much network support an assignment possibility must have for it to be included. Setting this to a higher value will mean fewer mistakes are made but at the expense of making fewer restraints. The “Strictness” setting relates to which residues may be involved in the local network to help resolve restraint possibilities. Considering only connections within a single residue makes few mistakes but doesn’t resolve much of the ambiguity. Using any residue makes most mistakes but most restraints. The default “Normal: Sequential residue support” is a useful compromise between these two extremes.
Residue and Chemical Shift Ranges
The “Residue Ranges” section is used to restrict the restraint generation to only specific regions of a molecular system. For example, this is useful if a protein is known to have an unstructured tail and the chemical shifts from this should not be considered because they would add unnecessary ambiguity. A residue range applies to one molecular chain and the peak dimensions on one side of a through-space correlation. This means that different dimensions can be restricted to different bits of a molecular system; useful with X-filtered NMR experiments. Residue ranges are applied to all forms of restraint generation, i.e. for assigned peaks, shift matching and network anchoring.
The “Chem Shift Ranges” tab is used to allow or disallow certain resonance signals from being used in the shift matching and network anchoring restraint generation. Different allowed ranges can be specified for each spectrum dimension (of the selected peak list) and a single dimension can have multiple ranges. In this way the user can control whether certain kind of resonances with distinct shifts are used, e.g. methyls. Often a ‘water notch’ is used, where two ranges with a small gap between are constructed to exclude peaks near the chemical shift of water; such peaks do not related to close macromolecule distances.
Caveats & Tips
Restraints are always made in a restraint list that is placed in the selected restraint set. By default a new restraint set is made, thus to make restraint lists that go together in an existing set the right one must be selected in the “Settings Tab”.
A new restraint set should always be used if any resonance to atom assignments have been changed, otherwise new restraints will still obey the original “fixed” assignments that were present in the old restraint set, not the new ones.
Peak List: Selects which peak list is used in the generation of distance restraints
Clone: Clone popup window
Help: Show popup help document
Close: Close popup
Documentation missing
Restraint Set: Selects the restraint set in which to make new restraint lists
Isotope Labelling: Selects the isotope labelling scheme, if any, to filter restraint possibilities
Shift Match Guide Structure: Selects a structure that can be used to remove distal resonance pairs from distance restraints
1.0: Threshold for peak figure-of-merit value, below which peaks will not be considered for making restraints
0.1: If using an isotope labelling scheme, the minimum proportion of spin active isotope to consider a resonance for restraints
20.0: The maximum allowed distance if filtering restraint possibilities using a structure
Intensity type: The kind of intensity to estimate distances with
Distance function: How to derive restraint distances from peak intensity
Only assigned ref intensities: Whether to consider only assigned peaks when calculating average, reference intensity
1.0: The peak intensity that corresponds to the reference distance. Defaults to peak list mean
0.2: The fractional lower restraint bound, defaults to 20% closer
1.72: The absolute lower distance limit, overriding fractional error
3.2: The distance corresponding to the reference peak intensity; to perform calibration
0.2: The fractional upper restraint bound bound, defaults to 20% further
8.0: The absolute upper distance limit, overriding fractional error
Documentation missing
Molecular system: Selects which molecular chain to set sequence/residue ranges for
Table 1 | |
Dimensions | Which spectrum dimensions the residue range filter applies to (Editable) |
Chain | The molecular chain of the allowed residue range (Editable) |
Start | The first allowable residue position to make restraints for (Editable) |
End | The last allowable residue position to make restraints for (Editable) |
Add Residue Range: Add a new allowable residue range to the table
Delete Residue Range: Remove the currently selected residue range
Documentation missing
Consider aliased shifts?: Whether peaks that are matched to shifts might be aliased one or more sweep withs; i.e. not at their real PPM
Ignore diagonal peaks?: Whether to avoid matching shifts to diagonal peaks
Copy From: Copy shift match tolerances settings from a different spectrum (overwrites old values)
Copy From entry: Copy shift match tolerances settings from a different spectrum (overwrites old values)
Table 2 | |
Dimension | The spectrum dimension that the tolerance applies to |
Min/default PPM tolerance | If a peak has no recorded linewidth, how far in ppm shift values can be from peak positions to make a restraint. With a linewidth, the absolute lower limit to this ppm difference (Editable) |
line width multiplier | A factor that governs how the peak-shift match tolerance scales with peak linewidth (Editable) |
Max PPM tolerance | If a peak has a recorded linewidth, the absolute upper limit to the peak-shift ppm difference (Editable) |
Documentation missing
Table 3 | |
Dimension | The spectrum dimension to which an allowed chemical shift range applies (Editable) |
Isotope | The isotope(s) relevant to the spectrum dimension |
Start | The lower bound of the allowed chemical shift range for matching shifts to peaks (Editable) |
End | The upper bound of the allowed chemical shift range for matching shifts to peaks (Editable) |
Add Chem Shift Range: Add a new chemical shift range for matching peak positions to chemical shifts
Delete Chem Shift Range: Remove the selected chemical shift range from the table
Documentation missing
1.0: The threshold NOE network sore, below which an assignment possibility will not be considered
Assign Peaks: Whether to assign the peaks of the input peak lists according to the filtered assignment possibilities
Strictness: Selects which kinds of NOE network information can filter assignment options;
higher strictness means fewer errors but fewer restraints
0.04: When generating restraint possibilities, the maximum difference between peak position and 1H chemical shift
0.4: When generating restraint possibilities, the maximum difference between peak position and 15N chemical shift
0.4: When generating restraint possibilities, the maximum difference between peak position and 13C chemical shift
Shift List: Selects which list of chemical shifts to compare to peak positions
Table 4 | |
Spectrum | The experiment: spectrum of the input peak list |
PeakList | Peak list serial number |
Use? | Whether to use this peak list in the NOE network and restraint generation (Editable) |
Num Peaks | NUmber of peaks in the peak list |
Make Assigned Restraints: Make restraints using the assigned peaks in the selected peak list, according to the settings
Make Shift Match Restraints: Make restraints by matching peaks in the selected list to chemical shifts
Test Shift Match: Make an analysis of peak to chemical shift matches, without committing any restraints
Network Anchor Restraints: Run the network anchoring calculation to generate restraints