The Pick & Assign module works beautifully with beautiful data, but if your data is more challenging, you may find you struggle at first. Here are a few ways to investigate your data and make changes to enable you to pick and assign your 3D spectra successfully.
Sometimes you may find that the peaks which are picked are in a neighbouring plane to the one you are currently looking at, or indeed in a plane even further away. This makes it difficult to assess quickly whether all the peaks were picked correctly. This may be a particular problem if your spectra have different digital resolutions in the z-dimension or are very highly resolved in the z-dimension. In this case it can be helpful to increase the number of planes shown in your display. You can easily increase the number of visible planes by incrementing the number to the right of your z-axis position. You may find that showing 3 or 5 planes simplifies the Pick & Assign process.
Sometimes it can be helpful to look at your 3D spectra a different way on to spot difficulties. Typically, you might be looking at a triple resonance spectrum with 15N in the z-dimension. It can be worth flipping it (with YZ) to show 13C in the z-dimension instead. This will allow you to check the 15N line shape. It may show that you need to extend or narrow the tolerances within which to pick your spectra. You can check what these tolerances are and change them by right-clicking on the gear icon in the Pick and Assign module. It is possible to set the tolerances separately for different specta.
You can also overlay your HSQC spectrum on the 3D when you have the 1H and 15N in the x-y plane (simply left-drag your HSQC from the side-bar into the new 3D spectrum display module). This allows you to check that the spectra are well calibrated. If they are not calibrated very well and there is an offset between the spectra, this will affect the program's ability to pick the 3D peak correctly. You can change the spectrum calibration by right-clicking and going to Calibrate Spectra or selecting a peak in each spectrum and right-clicking on one of the peaks and going to Calibrate Spectra from Peaks...
Sometimes the quality of your spectra may be such that you have to think a bit more creatively about how to work with them. Perhaps it is best to do the peak picking manually. You can then still use the Pick and Assign module to navigate to each strip and do the assignment (using Assign Selected), but you can pick the peaks manually. Or perhaps you can pick the peaks in some spectra automatically, but have a second 3D window in which to pick other spectra manually. This could be useful if some spectra have a particularly low signal-to-noise levels which means it is difficult to avoid picking noise peaks when doing the automatic peak picking.
If you are struggling with the quality of your CB-containing spectra, you could also decide only to pick and assign your CA-based spetra and only use these to link residues sequentially during the backbone assignment process. When the CA links are ambiguous, the CB-spectra can then be overlayed manually and used to determine the correct sequential link.