To assign HCCH experiments you should first ensure that you have entered an appropriate HCCH experiment in the parameters tab. Select this list in peak list chooser labeled hcchtocsy. Make sure you check that the pattern and tolerance information is set for the list. The proton dimensions should be set to i.h* and the carbon dimension to i.c*. Also, RunAbout needs to be able to determine which hydrogen dimension represents a hydrogen directly bonded to the carbon atom detected in the carbon dimension. To do this, open the Peak Inspector for the HCCH Tocsy list, choose Edit->Reference from the Peak menu, and enter a value into the relation field for that hydrogen dimension. For example, if your dimensions are H, H and C, and the second dimension is the hydrogen attached to the carbon (third dimension) enter D3 into the second relation field. You can also do this in the console: "nv_peak relation {} D3 {}".
The first step in the assignment process is to set up clusters of peaks that are connected to a common proton/proton pair. While this protocol is not strictly necessary, the clustering technique allows us to set up peaks that will be used to anchor the sidechain assignments for each residue. At present, the code is set up to cluster peaks based on the first (the directly detected proton) and second dimension (an indirectly detected proton) of the experimental data set. If your datasets are different you may need to have the script adjusted. Click the Cluster Peaks button in the HCCH Peaks section of the Actions Tab to actually carry out this process.
After the peaks are clustered, the clusters are also searched to find those that have a C13 ppm that is greater than 50 ppm, and thus are potential HA, CA pairs (excluding Gly residues). These clusters are stored into an internal array, sorted by the carbon chemical shift.
As with the backbone assignment mode, the next step after clustering is to examine, modify and possibly delete some clusters. This is done, as usual, in the Helm, which will now look like this:
The Navigation controls let you move back and forth through the list of clusters identified in the previous list. The chemical shift of the directly detected proton and its attached carbon are shown labeled as PPM C and PPM H. The list below shows the peak number and proton chemical shift for all the peaks in the cluster. One entry will have the shift shown in the PPM H entry, and is presumably the HA proton.
As you step through the cluster list the spectrum display will be updated to display the corresponding region of the dataset. The indirectly detected proton will be shown on the X axis, and the directly detected on the Y axis.
There are three classes of editing actions to do as you step through the clusters. First, for the convenience of our subsequent analysis every cluster strip should have a diagonal peak. Sometimes, for example for HA peaks near the water, the peak may be missing. RunAbout will draw a green horizontal line at the expected position of the diagonal. You will be alerted to the absence of a diagonal peak by the phrase No Diagonal appearing in the Helm display. If no peak is present you can add it by clicking the Add Diagonal Peak button. Second, some clusters correspond to just noise or artifacts and can be deleted by clicking the Delete Cluster button. Finally, you make changes to the set of peaks present in the cluster by using key bindings in the spectrum.
To add a peak that has not been picked, position the mouse cursor over the desired peak location, and press the a key. To delete a peak, select the peak (click on it with the cursor in selection mode) and press the d key. To remove a peak from the cluster, without actually deleting it, select the peak and press the r key.
Having worked through all the clusters you are now ready to assign the sidechain atoms. The assignment process works by stepping residue by residue through the sequence and evaluating possible assignments. RunAbout uses the chemical shifts of the previously assigned CA and CB atoms to present you with HCCH clusters that are likely related to that residue. Switch into HCCH Assignment mode to start the process. In this mode the RunAbout helm appears as in this figure.
The top section of the helm lets you navigate through the residues, by either clicking the up down arrows, or entering a number in the residue field (and hitting the Return key). The amino acid type of the current residue will be shown, and the Atom Assignments section will be updated to show the names and assigned chemical shifts of the assignable atoms in the residue. When you start a new residue, you will typically see chemical shifts for the ca, cb, n, and h atoms (as these were determined in the backbone phase).
The HCCH Peaks section of the helm will have three lists of peaks. At first, only the first list will be populated. It will contain a list of peaks and their proton shifts that are candidates for the assignment of the HA atom. The list of peaks consist of any HCCH peaks with a proton shift between 6.5 and 3.3 ppm, and whose carbon shift is close to that of the CA carbon of the current residue.
The list of possible peaks can be relatively long and therefore somewhat time consuming to evaluate. Fortunately, we have additional information that can be used to restrict the possible peaks. In particular, since we will generally also know the CB shift for the residue, we can limit our peaks to those that are in a cluster that have a peak with a carbon shift corrersponding to the CB shift. This filtering can be done by setting the "With HB Peak" button located above the peak lists. A less restrictive form of this filtering can be done by instead setting the "With HB Int." button. This will look, rather than for an already picked HB,CB peak, for intensity above a threshold in the dataset at the expected position of the HB,CB peak.
The spectrum display area shows three strips. As soon as you navigate to a residue the first strip will show the H-H view of the HCCH Tocsy at the carbon plane of the CA residue. The Y axis will show the full range of proton chemical shifts, and the X axis will be set to show an area that is likely to encompass any HA shift.
When you select (by clicking with the mouse) an entry in the ha list the display region of the first strip will be changed to narrow in on the X axis around the the chemical shift of the selected peak in the list. At the same time the second list (labeled protons) will be updated to show the peaks and shifts of all the protons in the cluster corresponding to the selected peak (one of the peaks in the list will be the same as the selected peak). These are all the possible assignments for the protons of that residue.
At this point you want to be checking whether the originally selected peak is reasonable for this residue by evaluating the list of protons. Clicking on any peak/proton pair in the protons list will perform several actions that will help you make this assessment. First, the chemical shift of the proton will be scored as a possible assignment for each proton in the residue. The calculated score will be a probability (from 0 to 1) based on the chemical shift distribution of atoms of that type (from BMRB statistics) for both the proton and the carbon shift selected in the third list (see below). Any probabilities above a threshold will have the probability value displayed in the right column of the atom assignments table, and if the probability is above 0.05 the atom row background will be set to green (rather than red). Second, the display of the second strip will be updated to display the region of the dataset corresponding to the selected peak. The x axis will be set to a narrow range around the proton chemical shift, and the carbon plane will be set to the carbon chemical shift of the peak. Both the first and second spectral strips will have green horizontal lines at the position of each proton in the cluster, and a green diagonal line intersecting the center at the position of the diagonal peak. Finally, a search will be done of the dataset to find carbon planes with proton peaks consistent with the protons in the cluster.
This search is done by scanning through the dataset and extracting a vector at each carbon chemical shift that corresponds to a proton shift of the selected proton. Each vector is compared to a reference vector formed from a slice at the shifts of the alpha carbon and proton. If the scanned vector corresponds to the same residue as that of the reference the vectors should be similar. That is, they should have peaks at similar proton shifts along the vector. A mathematical similarity measure is calculated between the pairs of scanned and reference vectors and any vectors with a value above a certain threshold will be listed in the third list (labeled carbon). Three numbers appear in each list entry, the carbon plane number, the corresponding carbon chemical shift, and the similarity measure. The similarity measure will range from zero up to the number of protons available for matching in the residue. The list will be sorted in descending order of the similarity measure, so the most likely carbon matches will be at the top. Typically you'll see several values with close numbers corresponding to adjacent carbon planes.
In the example shown below there is an HA proton selected in the first list, a proton selected in the second list, and a carbon in the third list. In this example, we've selected the HA proton at 4.403 ppm, another proton at 1.246 ppm, and a carbon at 21.6 ppm. The green highlighting and probability level of 0.92 indicate that these are consistent with the expected values for the gamma carbon and its attached methyl protons. The matching vector corresponding to plane 96 at 21.6 ppm is clearly better than the match at plane 51, and better than the nearby planes 94, 95, 97 and 98. Though not shown here, in this example selecting any other plane yields a lower value for the probability score as well.
The chemical shift values of these peaks can be assigned to the specific atoms by clicking on the buttons in the "Atom Assignments" table. Clicking the button labeled ha will take the proton chemical shift for the HA atom of this residue from the values in the selected row of the first (ha) list. Clicking the other buttons, hb or hg2 in this example, will take the proton shift for the corresponding atoms from the selected row in the second list, and the carbon shift from the selected row in the third list.
The carbon shifts listed correspond to the exact shift of each plane, and thus are only as precise as the digital resolution in that dimension. Peak positions are interpolated to have greater precision. You can use a peak position for the carbon dimension by selecting a peak (of the current residue) in the second spectral window prior to clicking the proton button. If no peaks are selected a confirmation dialog will be displayed asking whether or not to use the value of the carbon plane.
Additional information is always useful in confirming that one has the correct set of peaks assigned. One of the most useful complements to the HCCH experiments is the 15N-Noesy experiment. RunAbout is designed to allow you to simultaneously view this spectrum as you work through the HCCH spectra. To use this feature simply select the appropriate peak list on the n15noe line of the Parameters tab. Now each time you navigate to a residue the third spectrum will be updated to display the appropriate region of the NOE spectrum. One is generally looking for NOEs from the atoms of the current residue to the amide proton of the subsequent residue. Accordingly RunAbout will show the NOE spectrum at a plane corresponding to the amide nitrogen, and centered on the X axis at the amide proton, of the next residue. You can vary which residue is used by changing the "NOE Res" value. By default it is set to 1, but you can change it to any other value. Horizontal green lines will be drawn at the chemical shifts of the protons in the current reside so it is easy to identify the corresponding NOE peaks.
This figure illustrates the three spectral strips of the HCCH mode for a Val residue. The first strip is on the CA plane, and the second strip on the CG plane. The last strip is the NOESY spectrum and clearly shows the NOEs between the alpha, beta and gamma protons of the Val 17 to the amide proton Glu 18.
