Sparky helps you assign and integrate peaks in NMR spectra. The input is spectrum data in Felix matrix format or UCSF format. The output is ascii lists of assigned peaks, chemical shifts, volumes, linewidths, .... Finding peaks and making assignments is done manually through a graphical user interface. Sparky does not do processing or distance geometry, molecular dynamics, or make toast.
Assignment w1 w2 Volume lw1 (hz) lw2 (hz) G16H3'-H8 4.905 8.010 7.15e+06 ga 28.6 20.0 G16H4'-H8 4.439 8.013 5.42e+06 ga 35.3 16.9 T17H6-G16H8 7.205 8.004 1.68e+06 T17H7-G16H8 1.459 8.008 2.09e+07 ga 27.5 24.1 T17H2"-H1' 2.509 5.840 4.68e+07 ga 41.2 17.6 ...
The command to run Sparky is "sparky". The usual installation location is /usr/local/sparky and there is a link to the main program in /usr/local/bin. If either of these directories is in your search path you can start Sparky from any directory by typing "sparky".
Convert data processed with NMRPipe to UCSF format with the program pipe2ucsf that comes with Sparky. Data processed with Felix can be directly read by Sparky but the transmitter frequencies, spectral widths, reference shifts, ... in the Felix matrix must be correct. Set these with the Felix rmx command. Convert 2-D data processed with Varian's VNMR software using vnmr2ucsf that comes with Sparky.
Once you get frequency domain spectral data in Felix of UCSF format you can open it using "open" under the Sparky file menu. A new window will appear showing a contour plot.
The first thing you'll want to do is adjust the contour levels. Select "contour levels" from the tools menu. This will bring up a dialog that lets you set the lowest displayed level, number of levels, space between levels, and color of levels.
Press the "Apply" button to make your changes take effect. Dialogs in Sparky have an "Ok" button and an "Apply" button. Ok means make the changes and make the dialog disappear. Apply means make the changes and leave the dialog up. Since you pressed apply the contour dialog is still showing and you can make another change if you aren't satisfied with the new contour levels.
To show different regions of the spectrum move the scrollbars, resize windows, and use "zoom in", "zoom out", and "zoom previous" from the view menu. Note that the menu entries for zooming have 2-key accelerators listed next to them. The accelerator for zoom in is "zi". Typing zi in the spectrum window makes it zoom in. Another way to zoom in is to select the zoom pointer mode. There's a window of vertical buttons "select", "center", .... This window determines the pointer mode. Select the zoom pointer mode. Then drag a box in a spectrum window. The view zooms to show that region. By "drag a box" I mean press the left mouse button at one corner of the desired region, hold it down and move the pointer, and release the button.
You can show more than one view of a spectrum. Select "duplicate" under the view menu or type the accelerator "vd" in the window you want to duplicate. Another view will be created. You can adjust its contour levels and the region shown to be different from the original view. Note that the contour dialog has a menu of spectrum views so you can select which view you want to adjust. Each view has a name shown at the top of its window. The name is derived from the name of the spectrum with a number appended.
You can show as many views of as many spectra as you want. To open a view of a new spectrum use Open under the File menu. The views of any previously opened spectra remain displayed. To close a window double click on the button in the upper left corner of the window frame.
Now you've zoomed in to an appropriate magnification and set the contour levels to eliminate most of the noise. Choose the "find/add peak" pointer mode. If you click the left mouse button over a peak a marker will appear. You've placed a peak on the spectrum and it will appear in all views of this spectrum. You can also drag a box around a region, ie. press the left button and move the mouse and then release the button. A rectangular box is drawn and Sparky will place peak markers on all the peaks in the region. How does it do this? It uses a minimum height and a minimum linewidth. You set these using the peak dialog (accelerator "kt"). The linewidth is taken as the half height width.
If you want to move a peak marker switch to the "select" pointer mode and drag and drop the peaks. If you want to delete a peak switch to select mode, click on a single peak or drag a box around a region to select peaks. Selected peaks are outlined with a square. You can delete the selected peaks by pressing the delete key. If you delete some peaks by mistake you can use the "edit undo" command (accelerator eu) to restore them.
Select a single peak and bring up the assignment dialog (accelerator at). Type in a group name and atom name for each axis and press apply. Sparky is designed for assigning polymers -- proteins and nucleic acids. The group is meant to indicate the residue or base type and number (eg. P71, tyr36, a9, GUA15, ...). The atom names a specific atom of the group (eg. HN, CA, H2', H6, ...). Since Sparky does not interpret the group or atom names you can use whatever is useful for subsequent processing of peak lists. At the beginning of the assignment process you won't know much and can use group names like g1, g2, g3, ... and later rename them.
When you make an assignment a label is displayed near the peak. With the pointer in select mode you can drag the label to a new location.
Assignments can be displayed along the edges of a view. To show these resonance panels along the edges of a view use the "show resonances" command under the view menu (accelerator vR). The group and atom names will be shown along both axes. To see the alignment of peaks with these resonance names it help to display crosshairs, a horizontal and vertical line that follow the pointer. Bring up the view settings dialog (accelerator vt), and check crosshair display and press OK.
You've picked some peaks. Now try integrating one. Select the "integrate" pointer mode, then drag a box around a peak you want to integrate. The default integration technique is to use a Gaussian fit. If it worked, the peak will be recentered to the best fit position and will have a circle drawn around it. You can examine the fit by displaying one dimensional cross-sections along the edges of the view. Bring these slice panels up using "show slices" under the view menu (accelerator vS). The profiles of the spectrum data are shown in black and the peak fitting profiles in blue. The horizontal panel shows the horizontal cross-section at the pointer position.
To see the numerical volumes and linewidths bring up a peak list (accelerator lt). Press the "options" button in the peak list dialog and select volume and linewidth display and press OK. The list shows all peaks in the spectrum and can be written to a file by pressing the "save" button.
Peaks can be integrated using different methods. Gaussian or Lorentzian fitting can be done. Overlapped peaks can be simultaneously fit where all selected peaks within the lowest contour line are taken as a group. Groups can also include peaks not withing a common contour line but within a specified distance of each other. You can also integrate peaks by boxing or ellipsing. You drag a rectangle or ellipse around a peak and the heights at all interior points are summed up. The integration technique is chosen using the integration dialog (accelerator it).
Sparky maintains lists of resonances. By a resonance, I mean an atom name and a frequency. The frequency of a resonance is calculated from the positions of all the assigned peaks involving the atom. All peaks are weighted equally in the average. These resonances are listed in the assignment dialog to help you make assignments and can be shown along the edges of contour plots. Sometimes it is useful to maintain more than one set of resonances. If your study involves two similar molecules or spectra taken under different conditions (pH, temperature) that shift the resonance lines you probably don't want just one set of average resonance frequencies. To handle this Sparky associates a "molecule" name (eg. cyt-b5) and "condition" name (eg. ph7 25C) with each spectrum. These names are chosen by you and typed in using the spectrum dialog (accelerator st). Separate resonance frequencies are kept for each distinct molecule and condition.
To save the data you produce using Sparky use "save" under the "file" menu. This saves assignments, label positions, view window configurations, ..., jvfor a single spectrum. The file containing your processed NMR data (eg. noe150.ucsf) is never modified by Sparky. All Sparky data goes in a separate file in the ~/Sparky/Save directory (eg. named noe150.ses). Each spectrum has its own Sparky session file. When you are working with more than one spectrum, saving the data for each spectrum can be tedious. To save the data for all spectra use "project save" under the "file" menu. This creates a file in ~/Sparky/Projects which contains a list of the spectrum files you are working on. When you start Sparky at a later time you can open the project to load all your spectra. You can also load Sparky data for an individual spectrum by opening a file in ~/Sparky/Save in the same way you opened your processed NMR data. The open command recognizes both processed NMR data and Sparky data file formats.