3D peak picking (2)

3D peak picking and assigning

Instead of picking all the possible peaks in a 3D spectrum, we will use the peaks from the 2D 15N-HSQC to search the 3D spectra only around the HN root peaks. This procedure avoids picking of noise peaks, and at the same time provides a convenient way to perform the necessary manual inspection of the peak picking and assignment. The peaks picked in the 3D's will at the same time be assigned to the root resonances.

The 3D experiments that we will use here are:

• • Window HCN

  • HNCACB

    • • CA peaks (blue, positive): intra residue strong; sequential to (i-1) weak

      • CB peaks (red, negative): intra residue strong; sequential to (i-1) weak

  • CBCAcoNH

    • • CA and CB (blue, positive): both sequential to (i-1)

  • Window HHN

  • HNcaHA

    • • HA peaks (blue, positive): intra residue strong; sequential to (i-1) weak

  • HBHAcbcacoNH

    • • HA and HB peaks (cyan, positive): both sequential to (i-1)

Make sure that in the HCN window, both spectra are displayed; in the HHN window, HNcaHA and HBHAcbcacoNH are displayed, 15N-NOESY is not displayed (not active).

Click on M: Assignment: Pick & Assign From Roots. In Assignment: Pick & Assign From Roots that appears ensure that HN is selected in the "Root Window:" pulldown menu. Then select window HCN in the pulldown menu related to 'Target Windows' and click [Add Target Window:]. Do the same for the HHN window:

Now take a quick look at {Tolerances}, check if the parameters look OK (adjust the tolerances a bit), and go on to {Link Peaks}. In the table that shows up select the row corresponding to peak #106 (spin system {23}), which corresponds to a 15N-HSQC peak:

The selection of the peak has prompted the HN window to center on and mark the selected peak, and the HCN and HHN windows have navigated to- and centered on the HN coordinates of the 2D peak:

In the HCN window we see 4 peaks in the HNCACB:

• • 2 CA peaks (blue, positive): intra residue strong; sequential to (i-1) weak

  • 2 CB peaks (red, negative): intra residue strong; sequential to (i-1) weak

and 2 in the CBCAcoNH:

• • CA and CB (blue, positive): both sequential to (i-1)

The latter superimpose with the sequential peaks in the HNCACB. To inspect the peaks in both spectra, toggle the display of the spectra in the HCN window on and off (Figure 1).

In the HHN window we see 2 peaks in the HNcaHA:

• • 2 HA peaks (blue, positive): intra residue strong; sequential to (i-1) weak

and 2 in the HBHAcbcacoNH:

• • HA and HB (cyan, positive): both sequential to (i-1)

Of the latter, the sequential HA superimposes with the sequential peaks in the HNcaHA. To inspect the peaks in both spectra, toggle the display of the spectra in the HHN window on and off (Figure 1).

Figure 1.

Assignment note: Inspecting the chemical shifts of the HNCACB/CBCAcoNH peaks for this amide immediately identifies the residues in the sequence involved with high probablitity: The intra residual CA/CB shifts indicate that the residue is ALA, the sequential CA/CB shifts (i-1) indicate as the most likely candidates a VAL or PRO. Inspecting the primary sequence (M: Molecule: Residue Information, select Ala), shows that such a pair only occurs at Val34-Ala35. We therefore can confidently say that this is Ala35. However, for now, we don't assign it yet

Next, click [Pick & Assign Root Resonances] (make sure that for the HNCACB also the negative peaks will be picked by adapting the parameters in the Peaks: Peak Finding popup), the result is that for all the 3D spectra peaks have been picked and assigned to the same root frequency. After removing some noise peaks around the waterline in window HHN we have:

Toggle the spectra on and off and navigate +- 1 plane in the Nitrogen dimension to verify that all peaks have been picked properly, and no unwanted (noise) peaks remain.

This procedure needs to be repeated for all other NH root peaks by clicking on the rows in {Link Peaks}, or using the [Next root] and [Previous Root] buttons. To make life easy, you can use the option [Pick All & Assign Root Resonances] with appropriately set tolerances. This procedure works well for non overlapping peaks. However, overlapping peaks still need to be investigated manually, and sometimes peak positions have to be changed or manually picked. To avoid problems later on, I advice to either go through the complete list of roots one by one using [Pick & Assign Root Resonances], check that all peaks are correctly picked and assigned and modify the peaks and assignments where needed. Alternatively, pick and assign all peaks at once, and scan through the peaks later to remove unwanted peaks and add peaks where the peak picker failed. A well curated peak list will speed up procedures later on.

Note: The automated peak picking routine may pick peaks in nearby planes of the 3D spectra that are not displayed when inspecting the peaks at the precise Root Resonance. A way to search for these is to use the Resonance: Spin Systems popup, and display all the peaks in the spin system using the [Display Strips] and [Show Peaks] options. Check and modify the peak list where needed. In case a manual assignment is performed, the noise peaks can be deleted, and missed peaks can be added at a later stage. However, for automated assignment routines, it is crucial to have peak lists that are as free as possible from artifacts and contain all real peaks.

Previous: 3D peak picking (1)

Next: Assigning non-root resonances