Tools pulldown

FELIX includes several menu items that affect frequency-domain spectra in the workspace. Most of these functions are directly related to the transformation of multi-dimensional spectra, but several affect the processing of 1D data.


Tools/Buffers

Buffers are accessed from within the interface by selecting the Tools/Buffers menu item (<Alt>-tb).

Tools/Buffers/Store Work to Buffer

To store the current information in the workspace to a buffer, select the Tools/Buffers/Store Work to Buffer menu item (<Alt>-tbs) and enter the buffer number in the control panel. To visualize this information, you must change the stack depth to include that buffer. This action is useful for saving a spectrum when the workspace is needed for some other process.

Tools/Buffers/Load Work from Buffer

To load buffer information to the workspace, select the Tools/Buffers/Load Work from Buffer menu item (<Alt>-tbl) and enter the buffer number.

Tools/Buffers/Add Work to Buffer

The Tools/Buffers/Add Work to Buffer menu item (<Alt>-tba) adds the current contents of the workspace to the specified buffer. This action is especially useful for generating projections of multi-dimensional spectra and for co-adding the absorptive components of hypercomplex data.

Tools/Buffers/Multiply Work by Buffer

The Tools/Buffers/Multiply Work by Buffer menu item (<Alt>-tbm) can be used to multiply the data in the workspace by the contents of the defined buffer. This action is most often used to multiply the data in the workspace by an apodization function that was stored earlier in a buffer.

Tools/Buffers/Subtract Work from Buffer

The Tools/Buffers/Subtract Work from Buffer menu item (<Alt>-tbu) can be used to subtract the contents of the workspace from the defined buffer. This action is especially useful for generating projections of multi-dimensional spectra and for co-adding the absorptive components of hypercomplex data.

Tools/Buffers/Push Work to Stack Top

You can push data into the buffers by selecting the Tools/Buffers/Push Work to Stack Top menu item (<Alt>-tbp). This quickly stores the contents of the workspace on the top of the buffer stack. Every time you push onto the stack, you increase the stack depth by one.

Tools/Buffers/Pop Work from Stack Top

Used in conjunction with the Tools/Buffers/Push Work to Stack Top menu item, the Tools/Buffers/Pop Work from Stack Top menu item (<Alt>-tbt) is used to load the contents of the top of the buffer stack to the workspace and decrease the stack depth by one.

Tools/Buffers/Exchange Work/Stack Top

The Tools/Buffers/Exchange Work/Stack Top menu item (<Alt>-tbx) is used to exchange the contents of current workspace with the top of the buffer stack. This is useful for quickly moving data back and forth between workspace and the buffers.

Tools/Buffers/Zero Stack Depth

The Tools/Buffers/Zero Stack Depth menu item (<Alt>-tbz) is used to reset the stack display to show only the current workspace. This option is often used to bring the program back to a pre-defined state.


Tools/Lists

One of the most important features of the FELIX database facility is the ability to create fast read-only representations of database entities, or lists. Such lists do not actually hold database information per se, but consist of pointers to items in a database entity. Lists exist in the memory that is allocated to the 1D buffers, so extremely long lists may require memory reconfiguration.

The Tools/Lists (<Alt>-tl) menu item displays a pullright sub-menu of items, as described below.

Tools/Lists/List 1...4

The Tools/Lists/List 1...4 (<Alt>-tl1...4) menu item allows you to select from among four lists. All subsequent Lists menu items then operate on the selected list, which is symbolized by the pushed-in toggle button next to it. When you change the list number, cross-peak footprints referenced by that list are not automatically drawn.

Tools/Lists/Draw

The Tools/Lists/Draw (<Alt>-tld) menu item causes the contents of the current list to be displayed, using the color specified for that list.

Tools/Lists/Color

The Tools/Lists/Color (<Alt>-tlc) menu item allows you to specify a footprint color for each of the four lists. The default pen color aliases are listed in Table 7, page -148.

Tools/Lists/Zero

The Tools/Lists/Zero (<Alt>-tlz) menu item causes the current list to be re-initialized, or zeroed.

Lists are composed using any of the six menu items in the next group. The actions of all these menu items are cumulative, so you can use them to compile complex sets of cross peaks into one list.

Tools/Lists/Select Displayed

The Tools/Lists/Select Displayed (<Alt>-tlp) menu item adds all footprints in the currently displayed region to the current list.

Tools/Lists/Select Region

The Tools/Lists/Select Region (<Alt>-tlr) menu item adds only the footprints in the interactively defined region to the current list.

Tools/Lists/Select Line

The Tools/Lists/Select Line (<Alt>-tli) menu item adds footprints which intersect any part of the displayed cursor to the current list.

Tools/Lists/Find by Name

If the cross peak entity contains assigned peaks, then the Tools/Lists/Find by Name (<Alt>-tln) menu item may be used to create lists using peak names (using either partial or complete names).

Tools/Lists/Add One, /Remove One

Footprints may be interactively added and removed from the current list using the Tools/Lists/Add One (<Alt>-tla) and Tools/Lists/Remove One (<Alt>-tlo) menu items.

Lists may be organized and reviewed using any of the four menu items in the next category.

Tools/Lists/Merge Lists

The Tools/Lists/Merge Lists (<Alt>-tlm) menu item allows a third list to be derived from the union or the intersection of two other lists.

Tools/Lists/Sort

The Tools/Lists/Sort (<Alt>-tls) menu item sorts lists in ascending or descending order, according to the D1-center or D2-center fields in the cross-peak entity.

Tools/Lists/View

Items that are referenced in the current list can be reviewed using the Tools/Lists/View (<Alt>-tlv) menu item. A control panel allows you to specify whether to display peak centers, widths, and names using data points or ppm units. This displays a spreadsheet of cross peaks, which should not be edited.

Tools/Lists/Write

A hardcopy of the items that are referenced in the current list can be created using the Tools/Lists/Write (<Alt>-tlw) menu item. A control panel allows you to specify whether to display peak centers, widths, and names using data points or ppm units. These written lists are for record-keeping only and cannot be read back into FELIX.


Tools/Generate Spectrum/FID

The Tools/Generate Spectrum/FID menu item (<Alt>-ts) allows you to simulate spectra, FIDs, and noise.

The Spectrum from Parameters option generates single spectrum lines from an amplitude, frequency, and line widths. Lorentzian, Gaussian, and Voigt line shapes may be generated. The spectrum line can overwrite the workspace or be added to the workspace.

The FID from Parameters option generates a free induction decay (FID) from an amplitude, frequency, and a time constant. FIDs can overwrite the workspace or be added to the workspace.

The Add Noise option generates white noise. The noise can overwrite the workspace or be added to the workspace.

The Spectrum from Parameters option generates single spectrum lines from an amplitude, frequency, and line widths. Lorentzian, Gaussian, and Voigt line shapes may be generated. The spectrum line can overwrite the workspace or be added to the workspace.

The Spectrum from File option generates a 1D theoretical spectrum from a TurboNMR NMR shielding output file.

The Spectrum from Peaks option generates a 1D theoretical spectrum from a 1D peak entity in the database.


Tools/Functions

The Tools/Functions menu item (<Alt>-tf) displays a pullright sub-menu of items for dealing with the workspace and defining specific data values within the workspace.

Tools/Functions/Reduce to Real

The Tools/Functions/Reduce to Real menu item (<Alt>-tfr) is used to convert a complex spectrum to a real spectrum by discarding the imaginary part of the data in the workspace.

Tools/Functions/Complex

The Tools/Functions/Complex menu item (<Alt>-tfc) turns a real spectrum into a complex spectrum with a zeroed imaginary part.

Tools/Functions/Reverse

The Tools/Functions/Reverse menu item (<Alt>-tfv) causes the data in the workspace to be reversed by swapping datapoint values. Thus, data point 1 is swapped with data point N, and data point 2 is swapped with data point N-1, etc.

Tools/Functions/Complex Conjugate

The Tools/Functions/Complex Conjugate menu item (<Alt>-tfo) negates the imaginary part of the data in the workspace and has the effect of reversing a spectrum if it is done before a Fourier transform.

Tools/Functions/Magnitude Spectrum

The Tools/Functions/Magnitude Spectrum menu item (<Alt>-tfm) replaces the real part of the workspace with the square root of [(real)2+(imag)2], or the absolute magnitude of the data, and replaces the imaginary part of the workspace with the arctan (real/imag) or the phase array of the data, in the range -180 to +180°.

Tools/Functions/Power Spectrum

The Tools/Functions/Power Spectrum menu item (<Alt>-tfp) replaces the real part of the data in the workspace with [(real)2+(imag)2], or the power spectra, and sets the imaginary part of the data to zero.

Tools/Functions/Alternate Real/Imaginary

The Tools/Functions/Alternate Real/Imaginary menu item (<Alt>-tfa) turns a spectrum that has been separated into real and imaginary parts back into alternating real and imaginary parts. Alternating is the standard order where the real and imaginary parts of each data point are adjacent.

Tools/Functions/Separate Real/Imaginary

The Tools/Functions/Separate Real/Imaginary menu item (<Alt>-tfs) turns a spectrum that is in standard order into separate real and imaginary parts. In separate order, the real parts of all datapoints come first, followed by the imaginary parts of all datapoints.

Tools/Functions/Exchange Real/Imaginary

The Tools/Functions/Exchange Real/Imaginary menu item (<Alt>-tfe) causes the real and the imaginary parts of the workspace to be exchanged. This action is most often used when you need to add the real component of a FID (that is, the part of the real serial file) to the real component of a FID that is part of the imaginary serial file.

Tools/Functions/Shift Data

The Tools/Functions/Shift Data pullright (Alt-tfd) provides menu items to shift the data in the workspace a specified number of points to the left or right.

The Right Shift option shifts data to the right the specified number of points. The data shifted out is lost and zeroes come in on the other end.

The Left Shift option shifts data to the left the specified number of points. The data shifted out is lost and zeroes come in on the other end.

The Circular Right Shift option shifts data to the right the specified number of points. The data shifted out wraps around and comes back in on the other end.

The Circular Left Shift option shifts data to the left the specified number of points. The data shifted out wraps around and comes back in on the other end.

Tools/Functions/Set Data Size

The Tools/Functions/Set Data Size menu item (<Alt>-tfi) provides options to alter the size or number of datapoints in the workspace.

The Double Size option doubles the size of the data in the workspace by performing a linear interpolation between existing datapoints.

The Halve Size option halves the size of the workspace by averaging successive pairs of points.

The Set Size option sets the size of the spectrum in the workspace. When the new size is less than the previous size, the right end of the spectrum is thrown away. When the new size is greater than the previous size, the spectrum is padded with zeroed datapoints out to the new size.

Tools/Functions/Fold Data

The Tools/Functions/Fold Data menu item (<Alt>-tff) provides options for folding and unfolding data in the workspace.

The Fold Data symmetrizes the 1D workspace by co-adding the first and last points, the second and next-to-last points, etc., until a halved symmetrized 1D spectrum is created. Performing a point-generated fold on the workspace decreases the size by a factor of two. This action is convenient for nondiagonal symmetrization of 1D spectra.

The Unfold Data option re-symmetrizes the 1D workspace by regenerating the first and last points, the second and next-to-last points, etc., until a double symmetrized 1D spectrum is created. Performing a point-generated unfolding on the workspace increases the size by a factor of two.

The Low Point Fold option symmetrizes the 1D workspace by keeping the smallest of the first and last points, the smallest of the second and next-to-last points, etc., until a halved symmetrized 1D spectrum is created. Performing a low-point fold on the workspace decreases the size by a factor of two. This action is convenient for nondiagonal symmetrization of 1D spectra.


Tools/Mathematics

Tools/Mathematics/Set Data

The Tools/Mathematics/Set Data (<Alt>-tms) menu item includes three options:

The Workspace to Value option sets all points in the display spectrum to the values specified. Both real and imaginary points can be assigned in the control panel that appears.

The Point Range to Value option sets all points in the specified range to the values entered. Both real and imaginary points can be assigned in the control panel that appears. The range is specified in points.

The PPM Range to Value option sets all points in the specified range to the values entered, with the range specified in PPM.

Tools/Mathematics/Zero Data

The Tools/Mathematics/Zero Data (<Alt>-tmz) menu item contains several options for zeroing various datapoints in a spectrum:

The Zero Workspace option sets all points in the spectrum to zero. This zeros both the real and imaginary parts of every point.

The Zero Real option sets the real parts of all data points to zero. The imaginary parts are not changed.

The Zero Imaginary option sets the imaginary parts of all data points to zero. The real parts are not changed.

The Zero Greater Than option sets all points with values greater than the specified threshold to zero. This is a rather extreme way to wipe out large peaks in your spectrum and is sometimes used to remove a water peak.

The Zero Less Than option sets all points with values less than the specified threshold to zero. This is also a rather extreme action, used to wipe out baseline noise and small peaks.

Tools/Mathematics/Multiply Data

The Tools/Mathematics/Multiply Data menu item (<Alt>-tmm) multiplies all points in the workspace by a number. If the data in the workspace are complex, then the multiplier may be complex. This action allows you to change both the magnitude and phase of the data in the workspace.

Tools/Mathematics/Add To Data

The Tools/Mathematics/Add To Data menu item (<Alt>-tma) adds a number to all points in the workspace. The number may be real or complex.

Tools/Mathematics/Absolute Value of Data

The Tools/Mathematics/Absolute Value of Data menu item (<Alt>-tmv) replaces each point in the workspace with its absolute value.

Tools/Mathematics/Inverse of Data

The Tools/Mathematics/Inverse of Data menu item (<Alt>-tmi) replaces the data in the workspace vector with its inverse. This action takes the reciprocal of each point in the workspace (replace value by 1/value) and stores the new value in the workspace. Any zero points in the workspace are skipped to avoid a divide-by-zero error. This action can be used to create novel and interesting window functions for apodization.

Tools/Mathematics/Logarithm of Data

This menu item (<Alt>-tml) replaces each datapoint in the workspace with its natural (base e) logarithm. This action may be used to compute some novel window function for apodization. Any zero points in the workspace are skipped to avoid a divide-by-zero error.

Tools/Mathematics/Anti-Logarithm of Data

This menu item (<Alt>-tmn) replaces each data value in the workspace with exp(value), its natural (base e) anti-logarithm or exponential.

Tools/Mathematics/Derivative of Data

This menu item (<Alt>-tmd) takes the derivative of the data in the workspace and pushes it onto the current buffer stack.

Tools/Mathematics/Integral of Data

This menu item (<Alt>-tmt) takes the integral of the data in the workspace and pushes it onto the current buffer stack.


Peaks pulldown

Picking 1D or ND peaks or resonances is done with the items in the Peaks pulldown (<Alt>-k). The menu items in this pulldown work differently, depending on whether the active frame contains a 1D spectrum or an ND matrix.


Peaks/Pick One

This menu item (<Alt>-ka) allows you to select one peak at a time.


Peaks/Pick Region

To select peaks in a sub-region of your display, select the Peaks/Pick Region menu item (<Alt>-kg).

If you have a 1D spectrum or an ND spectrum for which the Pick Region Mode in the Preference/Pick Parameters control panel is set to Define by Cursor, this creates a small crosshair cursor that you can use to drag out a region.

Otherwise for ND spectra, FELIX picks-peaks in the displayed region (Pick Region Mode was set to Displayed Region), or you can specify a region through a control panel (if Pick Region Mode was set to Define via Dialog).


Peaks/Pick All

To automatically pick peaks in the full spectrum, select the Peaks/Pick All menu item (<Alt>-kp). This opens a control panel where you can set the peak-pick parameters (which also can be accessed through Preference/Pick Parameters).


Peaks/Remove One

This menu item (<Alt>-kn) deletes peaks one by one using the cursor.


Peaks/Remove Region

This menu item (<Alt>-kr) deletes peaks in a region that you select by dragging out with the small crosshair cursor.


Peaks/Remove All

This menu item (<Alt>-kl) deletes the peak entity.


Peaks/Edit

This menu item (<Alt>-ke) allows you to interactively adjust the cross-peak shape of a selected cross peak. Initially, this function allows you to interactively (that is, using the cursor) select a cross peak for editing. FELIX signals that this function is complete by changing the cross peak's footprint color. You may adjust the footprint position by clicking near the center of the footprint and dragging. You may adjust the cross-peak shape by clicking near the edge of the footprint and then dragging. This function remains active until you click a point in the display at which no cross-peak footprint exists.

Note:
This is available for ND peaks.


Peaks/Filter

The Peaks/Filter (<Alt>-pf) menu item includes a set of tools for use after peak picking. These tools act as filters, eliminating a subset of cross peaks based on the criterion you select. Different numbers of tools are available, depending on the dimensionality of the spectrum.

The Remove Redundant Peaks menu item (<Alt>-kf) removes redundant peaks for the set of picked 1D peaks.

The Remove Diagonal Peaks menu item asks you for a filter tolerance (in datapoints, ppm, or Hz) and then removes cross-peak footprints that lie (within the specified tolerance) on the diagonal (D1=D2). For a 3D data set this diagonal can be any of the plane diagonals (D1=D2, D2=D3, or D1=D3) or the body diagonal (D1=D2=D3).

The Symmetrize Spectrum menu item asks you for a filter tolerance (in datapoints, ppm, or Hz) and then removes cross-peak footprints that exist on only one side of the diagonal (that is, only those footprints at (D1,D2) which have symmetry-related partners at (D2,D1) are retained). This menu item also works for 3D spectra.

The Merge Multiplets menu item asks you for a filter tolerance (in data points, ppm, or Hz) and then combines footprints that have D1/D2 centers within the specified tolerance of one another. The resulting footprints have widths that represent the largest of the widths of the unfiltered footprints in each dimension.

The Filter by Width menu item asks you for a minimum and maximum footprint halfwidth (in datapoints, ppm, or Hz) and then removes footprints that fall outside the specified halfwidth limits.

The Filter by Volume menu item asks you for a volume threshold and then removes peaks that have volumes below the specified threshold. Before you can use this menu item you need to measure the volumes.

The Filter by Intensity menu item asks you for a lower and upper intensity threshold and then removes cross peaks that fall either outside or inside the limits.

The Remove Peaks in Region menu item asks you for a region (in data points, ppm, or Hz) and then removes footprints that fall inside the specified limits.

The Remove Redundant Peaks menu item removes footprints that have D1 and D2 centers at precisely the same position. The resulting footprints have widths that represent the largest of the widths of the unfiltered footprints in each dimension.

The Clean-up Ridges menu item asks you for a strip width (in data points, ppm, or Hz) and for the maximum number of peaks that there should be along each strip, and then removes the excess low-intensity cross peaks in each strip. This action can be used to clean up noise peaks when you know that, at a given frequency (within a certain tolerance), there cannot be more than a certain number of peaks (for example, in an HCCH-TOCSY spectrum along the aliphatic H dimension within 0.1 ppm in C and 0.02 ppm in H, there cannot be more than 10 peaks).

The Filter by Spectrum menu item cleans up the current spectrum. You can, for example, clean up a 3D HSQC-NOESY spectrum based on an HSQC spectrum. The function asks you for a spectrum and peak table to use (usually a 2D HSQC spectrum). Then you should specify which dimensions from the 2D to use for accepting peaks in the current spectrum within the given tolerance. All other peaks (most likely noise peaks) are deleted.


Peaks/Optimize

This menu item (<Alt>-kz) offers a powerful line-fitting interface for deconvolution of complex spectra into individual peaks, which are described by an analytic function of intensity, linewidth, and frequency. These functions allow precise integration of individual peaks.

Depending on the dimensionality of the current spectrum, two different menu items can be accessed:

Select the Peaks/Optimize/Optimize menu item to access the 1D line-fitting menu item and other menu items that display the actual data, the synthetic data, and residual data - either separately or all together in an overlay.

The interface for this funtionality consists of a display area and a control panel that provide convenient tools for picking peaks, automatically optimizing peak parameters, and determining peak integral values. The display provides essential graphical information by showing the real and synthetic spectra and additional information at all times. The control panels enable you to both manually and automatically adjust the peak parameters to best match the synthetic spectrum to the real spectrum.

To fit a spectrum, select the Optimize item in the Peaks/Optimize menu. The interface appears in the current graphics frame with your spectrum in the display area and beneath it the control panel, which contains buttons and sliders that permit you to fit the spectrum.

The Peak buttons let you select peaks, add and delete individual peaks, and set the peak-picker threshold. Whenever you pick peaks, the picker determines initial values for each peak's center, height, and width. After you pick peaks, there appear in the display, in addition to your spectrum (which is drawn in white), the synthetic spectrum drawn in red and a selected peak drawn in yellow. A blue dot at the top of each peak identifies the location of each picked peak and a yellow dot identifies the selected peak. You can change the selected peak by shifting to an adjacent peak using the buttons labeled Next and Previous or by using the mouse to pick the peak you want to select. The slider and readouts in the right half of the control panel allow you to modify the center, height, and width of the selected peak. As you adjust the slider, the display area switches to the fast-update real- time mode that shows the synthetic spectrum and the real spectrum, so that you instantly see how your changes affect the synthetic spectrum.

Selecting the Setup button activates another control panel that lets you set certain initial parameters such as the Peak shape, which can be Lorentzian or Gaussian, and optimization locks, which prevent the automatic optimizer from adjusting the specified parameters. These settings affect only the peak picker, so you must set them before picking peaks.

In this control panel you also can select the optimization function, its control parameters, and the length of the peak tails used to calculate the spectrum. You have the choice of three optimization methods: simplex, quasi-Newton, and simulated annealing. We find that the simulated annealing method performs the best, so it is the default method. The peak tails are given as a percentage of the peak height so, for example, a peak tails setting of 0.01 means that the peaks are calculated from their centers to the point where they fall to 1% of their maximum value. This setting is used to accelerate the optimization by avoiding needless calculation of the contribution of peaks in regions where they have little or no significant amplitude.

You can change the peak functions or parameter locks of a selected peak using the toggle buttons in the middle of the panel.

The Optimize/Threshold button provides a crosshair that sets the level under which the optimizer ignores the difference between the real spectrum and the synthetic spectrum. That is, the optimizer does not attempt to match the real and synthetic spectrum beneath this threshold. You can use the threshold to make the optimizer ignore baseline noise and accelerate optimization.

The Start button starts the automatic optimizer and the Stop button (the Start button changes to Stop after you start the optimization) provides a way to interrupt the optimizer during its operation. During the optimization, the spectra are updated in the display area to show the progress, and a numeric value related to the difference between the real and synthetic spectra is continually updated in the relative Error readout.

You can read out the integral values of each peak at the Integral readout and can set a normalization reference peak and multiplier with the Normal button.

On exiting the curve-fitting interface, the optimized peak data are stored in the existing 1D peak-picking entity.

Overlays plots the actual, synthesized, and residual data at the same time.

Undo Overlays removes the three data types mentioned above and displays only the actual data.

Actual Data plots only the actual data.

Synthesize Data plots the synthesized data only.

Residual Data plots the residual data only.

Save Current Data saves the current spectrum to a 1D disk file.

ND peak optimization

This menu is used for ND cross-peak footprint optimization and modeling of cross-peak shapes and volumes. All the menu items require that you already have an entity of picked peaks and an entity of measured volumes for at least one mixing time.

At the heart of these peak-optimization and -modeling functions is the notion of a model cross peak. Whereas an actual cross peak is described by a collection of adjacent data points in a matrix having intensities greater than the neighboring region, a cross-peak footprint is described by a set of center position and halfwidth values, and a volume is described as the total intensity of the cross peak's datapoints that lie inside the footprint; a model cross peak is defined as the distribution of intensity described by an idealized peak created from an analytic function of line widths, volume, and center positions. At present, FELIX uses either a Gaussian or Lorentzian line-shape function. The practical implication of these definitions is that FELIX can model the intensity of every cross peak that you have picked and measured a volume for, with a Gaussian or Lorentzian peak shape derived from the footprint and volume information.

The primary utility of these model cross peaks is twofold. First, you can display plots of model data or residual data (real minus model) to help with your early analysis and assignment work. There are many times when visually subtracting out a few well resolved peaks can reveal additional peaks hidden underneath. The second use for model cross peaks is to perform nonlinear least-squares optimizations on cross-peak shapes and intensities to yield the "best fit" values for peak center positions, halfwidths, and volumes; because the optimization algorithms seek to minimize the difference between the real data and the model representation of that data. Optimization can increase the confidence for assignment decisions based on peak alignment and improve volume buildup rate estimates for making distance restraints.

The Optimize menu item lets you use a conjugate-gradient minimization algorithm to improve the values for peak centers, halfwidths, and/or volumes that are stored in the peak and volume entities. The algorithm used is the quasi-Newton, as mentioned in the 1D line-fitting section. A control panel prompts you for a peak entity, a volume entity, a volume slot number, and the type of line shape to use. It also lets you select which values to optimize and which values to hold constant.

Keep in mind that, for each peak, there are two (or three or four) centers, two (or three or four) widths, and one volume; and that all are fit against a rather small set of data points (the points inside that peak's footprint). Accordingly, the fewer the values that are optimized per peak, the better-determined the algorithm (and the faster it executes). You may find that optimizing just volumes, then just centers, and so on, gives better results than optimizing everything at once. Again, this is primarily a concern when there are relatively few datapoints per footprint.

Once you have selected which values to optimize, select OK. FELIX reports how many sets of peaks it will operate on and then announces each set optimized and colors those peaks in green as it completes each set. When all sets are done, FELIX reports the initial and final penalty values in the text window. These penalty units are on an arbitrary scale, yet they represent the RMS deviation in the intensity values at every datapoint in the optimized region of the matrix that has any cross peaks. Lower penalty values correspond to a better fit.

The new values for centers, halfwidths, and volumes are stored in their respective entities. The results of this action cannot be undone, so you may want to save backup files of your peaks and volumes first.

Data modeling

So far we have only discussed the notion of actual matrix data. When you do a plot, the actual datapoints in the matrix are what get fed to the plot routine. In addition to actual data, FELIX can also display and analyze model and residual data.

The Actual Data menu item instructs FELIX to use the actual matrix data. This has the effect of undoing the actions of the Model Data and Residual Data menu items.

The Model Data menu item instructs FELIX to synthesize data from the current peaks and volumes, instead of reading datapoints from the matrix. All subsequent plots and analyses are performed on the synthesized model data. To return to using the actual matrix data, select the Peaks/Optimize/Actual Data menu item.

The Residual Data menu item instructs FELIX to use a blend of real datapoints from the matrix and synthesized data from the current peaks and volumes. Specifically, the blend is "one part real minus one part model"; most commonly called the residual. All subsequent plots and analysis are performed on this residual data. This feature can be used to measure the residual volume inside peak footprints that cannot be attributed to the peak, which is a sigma or uncertainty measurement for the real peak volume. To return to using the actual matrix data, select the Peaks/Optimize Peaks/Actual Data menu item.

The Model Parameters menu item can be used to select whether Gaussian or Lorentzian line shapes are used in the modeling process.

Note:
Be sure you remember to explicitly return to using actual data when you are finished investigating model and residual data.


Peaks/Brother Peak

The Peaks/Brother Peak menu item (<Alt>-kb) allows you to explicitly extend assignments to other cross peaks along the D1 or D2 dimensions. You first select, with the crosshair cursor, a cross peak that already has assignments in one or more dimensions. You then explicitly select cross peaks that share the same D1 or D2 assignments.


Peaks/List

The Peaks/List (<Alt>-kl or <Ctrl>-l) menu item serves as a quick interactive check of the status of individual peaks (chemical shifts, widths, peak assignments, and frequency assignments), which are listed in the text window or zoomed on in the peak table if it is open.


Peaks/Find

The Peaks/Find (<Alt>-kd) menu item can be used to find a specific cross peak. You can search for a cross peak by its assignment name or its number. Alternatively, you can use the List/Name menu item to create a list of cross peaks that match a search criterion.

If you choose to search the cross-peak entity for a specified assignment name, or partial name, FELIX creates a list that contains the discovered peak(s). A wildcard character (*) may be used as part of the assignment name. If a peak is found, the display expands to show the found peak or the peak changes color, depending on your choice.


Peaks/Name One Peak

Use the Peaks/Name One Peak (<Alt>+ko) command to select a peak interactively and then assign names to the peak.


Peaks/Example Peak

For information on this command refer to "Preference/Pick Parameters" on page 155, in particular the information there for Stella Peak Picking.


Measure pulldown


Measure/Cursor Position

The Measure pulldown includes functions for obtaining point numbers, ppm values, and corresponding data values. When you select the Measure/Cursor Position menu item (<Alt>-up) while a 1D spectrum is displayed, a vertical half-cursor appears. The current axis position and data value are reported as you move the cursor. The axis position is in axis-based units - if your axis is in points, it tracks in points, if your axis is set to ppm, it tracks in ppm. To quit the cursor-tracking mode, press <Esc>. The data value shown is the actual data value stored at that location in the workspace.

For ND spectra, accurate peak positions and heights can be read interactively from the display using this menu item. When you select the Measure/Cursor Position menu item, FELIX displays a crosshair cursor and reports the cursor position in x- and y-axis units (possibly the third and fourth dimension, as well); FELIX also reports the matrix value at that point. The information is continuously updated as you move the cursor, until you quit the cursor-tracking mode by pressing <Esc>.


Measure/Correlated Cursors

The Measure/Correlated Cursors (<Alt>-uc) menu item initiates multiple-cursor mode, which is used to correlate peaks in more than one simultaneously displayed plot. For example, you may plot a TOCSY, COSY, and NOESY spectrum in three different frames. When you select the Measure/Correlated Cursors menu item and move the cursor into one of the frames, the cursor becomes a crosshair cursor and crosshair cursors appear and track the identical positions in the other two frames. If the matrices are referenced and displayed with axis units of ppm, the cursors can be used to correlate peaks in the three spectra. Pressing the mouse button reports the current position. To quit multiple-cursor mode, press <Esc>.


Measure/Distance/Separation

For a 1D spectrum, in addition to spectrum positions and intensities, you can calculate the separation between any two spectrum features in a display with the Measure/Distance/Separation menu item (<Alt>-ud). You use a crosshair cursor to select two locations on the display. FELIX reports the separation in points, ppm, and Hertz. To exit, press <Esc>.

For an ND spectrum, the Measure/Distance menu item allows you to find the distance between the two atoms defining a particular peak (either through peak assignment or through frequency assignment) by clicking the peak. The distance is measured in the currently active molecule.

If the displayed object is a molecule (in Model mode only) then you can click on two atoms to obtain the distance between them in angstroms.


Measure/Integral/Volume

1D Integral

FELIX allows you to integrate the entire spectrum as a single integral or as shorter segments. To integrate the entire spectrum, simply select the View/Draw Integrals menu item (<Alt>-vs). If you want further options dealing with integrals you can use the Measure/Integral/Volume menu item (see below).

You need to define integral segments by selecting the Add Segment option. Integral segments are added by dragging out a segment region with the mouse. As long as you select a valid region within your spectrum, you can add additional segments without re-selecting the Add Segment option. To exit this mode, click outside the spectrum. An example of a sequential integral spectrum is shown in the following figure.

If you make a mistake while selecting individual segments or if you want to modify the current list of segments, you may delete a small subregion of segments graphically. Select the Remove Segment option to create a small crosshair cursor. Then drag out a region of segments to delete.

To delete all the segments, select the Remove All Segments option in the Measure/Integral/Volume menu item. This deletes the current integral segments entity from the database and requires confirmation via a dialog box.

For some spectra, it is impossible to accurately define baseline points. By selecting the Slope and Bias option, you can adjust these parameters in real time using dials and buttons. However, if the baseline is significantly distorted, adjusting the slope and bias may not be able to generate correct integral shapes. By adjusting the slope and bias, you can set an integral value to anything you want. So use these adjustments cautiously.

The Slope slider adjusts the integral slope parameters between -1 and 1. The Bias slider adjusts the integral bias parameters between -1 and 1.

ND volume measurement

FELIX calculates the volume of a cross peak as the integral of all data-point intensities inside the cross-peak footprint. The reserved symbol hafwid controls the relative size of all footprints, affecting the resultant volume measurements (please see the online FELIX Command Language Reference Guide for more detailed information about this symbol). Because the peak picker determines the cross-peak footprint widths from the peak's half-width at half-height, tall peaks that have better signal-to-noise ratios have relatively smaller footprints than small peaks with worse signal-to-noise and thus relatively larger footprints. In this way, the volume of a strong peak is the sum of relatively few data points of high intensity, while the volume of a weak peak is the sum of relatively many data points of low intensity.

The Show One Volume option calculates and displays the volume of a single cross peak. The arrow cursor becomes a large crosshair, which you use to click the peak of interest. The volume for the selected cross peak is calculated and displayed in the text window. When there is a current volume entity, both the raw calculated volume and the stored volume from the volume entity are displayed. As long as a peak is selected, the action repeats itself. To end the action, click in a region of no peaks or press <Esc>.

The Measure All Volumes option measures the volumes of all cross peaks. A control panel asks you for the cross peak entity name, the volume entity name, a slot number, and the mixing time for the current matrix in seconds. If the named volume entity does not yet exist, you are asked for the total number of slots to be built into the new volume entity. When the volume entity does exist, FELIX first checks to make sure that the number and IDs of all cross peaks match the number and IDs of all volumes. This assures that the two entities are compatible. Specifying a slot number that you have already used overwrites those volumes.

The measured volumes depend on the center and width of each cross peak and the reserved symbol hafwid, which can be adjusted using the Preference/Peak Display menu item. Measuring volumes is very fast for small cross-peak sets, but can take tens of seconds for very large numbers of cross peaks.

The Measure Buildups option is different from the Measure All Volumes option in that it measures the volumes not only in one spectrum but in a series of spectra within the same buildup and stores the result in the consecutive slots of the same volume entity.

The Remeasure Buildup for One Peak option remeasures a buildup for a particular peak.

The Delete Volumes option deletes the current volume entity, after asking you for confirmation. There is no way to delete a portion of the volumes.


Measure/Buildup

Measure/Buildup/Show Buildup

The Measure/Buildup/Show Buildup (<Alt>-ubs) menu item gives a graphical representation of all the stored volume slots for a selected cross peak. Select a cross peak with the large crosshair cursor. The action makes a 2D graph of volume versus time, showing the volume buildup for that cross peak. The initial datapoint of zero volume at time zero is explicitly included.

Measure/Buildup/Fit Buildup

The Measure/Buildup/Fit Buildup (<Alt>-ubf) menu item allows you to fit the buildup data to any of seven different functions.


Measure/J Coupling

FELIX provides a set of tools for extracting J-coupling constants from a variety of spectra. These coupling constants then can be used to calculate torsion angle restraints for structure determination studies.

Because of the complexities of overlapping COSY peaks and multiplet "splitting" effects within any one peak, you can calculate the J-coupling constants for only one peak at a time. Based on the quality of the results, you are then free to add each J-coupling measurement to the J-coupling entity one at a time.

Measure/J Coupling /DQF

The Measure/J Coupling/DQF (<Alt>-ujd) menu item uses a sophisticated line-fitting algorithm to calculate the true centers of each lobe of a DQF-COSY multiplet and then measures the separation in Hertz. This menu item works only on non-overlapping peaks with four primary lobes (down, up, up, down). It cannot robustly handle peaks with more lobes due to "splitting". You use a crosshair cursor to select a peak. The function then calculates and reports the J coupling and sigma to the text window.

The algorithm works as follows: for each dimension, FELIX divides the cross-peak footprint in half and then projects each half (sum the points) down to a 1D line segment. This yields two 1D line segments that each represent the line shape of two lobes of the cross peak (one up, one down). The number of datapoints in each line segment depends on the size of the cross-peak footprint in that dimension. Next, each of these line segments is peak picked (to yield two peaks - one positive, one negative) and then passed through the 1D curve fitter to optimize the two peak centers, widths, and heights to best-fit Gaussian line shape models. This optimization step is responsible for finding the "true" centers from the "apparent" centers given by the peak-pick routine, for each line segment. This yields two independent measurements for the separation in each dimension. The function reports the average separation as the J-coupling value and reports the deviation from that average as the sigma, or uncertainty, of that J-coupling value, for each dimension.

Measure/J Coupling/Manual ECOSY

This menu item (<Alt>-ujy) allows you to measure a coupling between two subpeaks of a typically ECOSY spectrum. Once you use the cursor to define the boundaries of two subpeaks, FELIX uses the optimizer (similar to DQF) to measure the distances.

Measure/J Coupling/Heteronuclear ECOSY

This menu item (<Alt>-uje) measures J couplings in heteronuclear ECOSY-type 2D experiments (Griesenger et al. 1986).

Measure/J Coupling/heteronuclear FIDS

This menu item (<Alt>-ujf) measures a 3J coupling on two HSQC spectra using the FIDs (fitting of doublets and singlets), where you measure the fully decoupled and a partially coupled HSQC spectrum. After peak picking, the coupling constants can be extracted using time-domain fitting (Schwalbe et al. 1993).

Measure/J Coupling/Heteronuclear FIDS/ECOSY

This menu item (<Alt>-ujs) measures the 2JCN and 3JCN coupling constants on three 2D HSQC experiments, using the combined C1-FIDS or FIDS-ECOSY method (Rexroth et al. 1995a).

Measure/J Coupling/heteronuclear DQ/ZQ

This menu item (<Alt>-ujz) measures the J couplings on a pair of double-quantum and zero-quantum 2D experiments like HN(CO)CA or HNCA (Rexroth et al. 1995b).

Measure/J Coupling/Heteronuclear 3D ECOSY

This menu item (<Alt>-uj3) measures the J couplings in a 3D ECOSY-type experiment.

Measure/J Coupling/HSQC-J

This menu item (<Alt>-kjh) allows you to measure the J coupling via a series of HSQC experiments, finding the coupling constant by interpolating the zero crossing of the volume series. Thus you must have picked peaks in a series of HSQC spectra and have measured volumes measured and stored them in the volume entity.

Measure/J Coupling/Manual Separation

The Measure/J Coupling /Manual Separation (<Alt>-ujm) menu item is a primitive tool, compared to the line-fitting algorithm described above. You select a peak with a large crosshair cursor. The chosen peak is then plotted in an expanded blown-up frame to represent as much peak shape as possible. You then use a small crosshair to drag out a rectangular shape. You should try to align the four corners of the rectangle on the true centers of the four lobes of the peak. The menu item then calculates the separation in Hertz from your cursor corners and reports the J-coupling values in the text window. There are no sigma terms with this method, since there is only one measurement for each dimension.

Measure/J Coupling/Volume Ratio

The Measure/J Coupling/Volume Ratio menu item (<Alt>-ujv) displays the volumes of two peaks (e.g., an off-diagonal and a diagonal) selected via the cursor from the currently displayed ND spectrum (usually triple resonance). You can then use this volume ratio for calculating J coupling with an external program.


Measure/Relaxation

The Measure/Relaxation menu offers a suite of tools that allow you to analyze a series of heteronuclear 2D relaxation spectra. It is assumed that peaks have been picked in one of the spectra or in a similar one with the same spectral widths.

Measure/Relaxation/Measure Heights/Volumes

This menu item (<Alt>-uv) evaluates peak heights or volumes in the series of spectra and has optimization features that can accommodate slight peak displacements relative to the initial peak table and moderate peak overlaps.

Measure/Relaxation/S/N Ratio

This menu item (<Alt>-us) determines the signal-to-noise ratio by analyzing one or more duplicate spectra and extrapolating to the remaining time points.

Measure/Relaxation/View Timecourse via Cursor

This menu item (<Alt>-uc) allows you to point at a peak in the displayed 2D spectrum and displays the series of peak heights or volumes of that peak with error bars as determined by the previous two menu options. The best-fit exponential decay curve is also displayed if one has been fitted to the data, and the relaxation rate is displayed in the status bar and the text window.

Measure/Relaxation/View Timecourse via Item

This menu item (<Alt>-ut) lets you choose a peak number whose time course you want displayed.

Measure/Relaxation/Fit R1/R2/NOE

This menu item (<Alt>-uf) analyzes the peak height or peak volume data. R1 and R2 timecourses are fitted to appropriate exponential decay curves, taking into account the experimental uncertainties. The resulting relaxation rates are displayed in the text window and are stored in a database table. The database table also holds the other parameters of the fitted curve in order to reconstruct it if necessary. The heteronuclear NOE is evaluated by taking the ratio of the peak heights in spectra acquired with 1H saturation and without 1H saturation. Signal-to-noise evaluation is built into this procedure by analyzing a second pair of spectra.

Measure/Relaxation/Modelfree input

This menu item (<Alt>-um) extracts the relevant parameters from the FELIX database and prepares a rudimentary input file for the ModelFree program of A.Palmer (available at
http://cpmcnet.columbia.edu/dept/gsas/biochem/labs/palmer/).


Measure/Scalar/Normalize

FELIX provides the ability to normalize the integral of any segment of the spectrum to an arbitrary value. Four normalization menu items are available under the Measure/Scalar/Normalize menu item (<Alt>-un). After normalization, the volume element in the integral segment entity is updated to the normalized value.

By Item Number of Segment

This menu item creates a list box where you graphically select the segment to normalize based on its beginning and ending point. You must also give a normalization value for this segment.

By Data Point Limits

This menu item opens a control panel in which you enter a low and high point to define a normalization range, as well as the normalization value.

Select Segment via Cursor

This menu item generates a small crosshair cursor that lets you select the segment to normalize by dragging to enclose the segment.

Raw Absolute Integrals

This menu item stores and displays each segment's integral as its raw intensity, with no normalization at all.

The rate of volume buildup for a cross peak in a set of NOESY matrices is related to the distance between the two corresponding atoms in the molecule of study. Prior to deriving distance restraints from a set of volume buildup rates, it is necessary to define a scaling constant taken from cross peaks that correspond to fixed interatomic distances in the molecule. FELIX allows you to define a small set of reference cross peaks that all correspond to a single fixed distance. The buildup rates of these peaks are then averaged to determine the scaling constant used to convert volume buildup rates into distance restraints. Due to the method of averaging these rates, it is very important that all cross peaks selected as scalar peaks correspond to one single interatomic distance in angstroms. The only exception to this rule is when the Empirical Fit method of restraint calculation is used. Here, the scalar entity consists of a series of scalar peaks that correspond to a range of interatomic distances.

We cannot overemphasize the importance of these reference peaks: they are crucial in defining NOE distance restraints.

Add One

The Measure/Scalar/Normalize/Add One menu item adds one more cross peak to the scalar peak entity. A control panel prompts you for the assignment names in D1 and D2 and a distance in angstroms. The program then searches the current cross-peak entity for a peak with those names, verifies that peak is not already a scalar, and then adds one more entry to the scalar-peaks entity. Remember that all scalar peaks should represent one single distance. Basing the scaling constant on an average of different distances is not valid. The only exception to this rule is when the Empirical Fit method of restraint calculation is used (see above).

Add One via Cursor

The Measure/Scalar/Normalize/Add One via Cursor menu item is similar to the previous Add One menu item, but you can add the peak by selecting it with the cursor.

Delete One

The Measure/Scalar/Delete One menu item lets you remove one peak from the scalar entity. A control panel shows you a list of all scalar peaks, by assignment name, and allows you to select one scalar to remove from the entity. You may repeat this action to delete several scalar peaks.

Clear All

The Measure/Scalar/Clear All menu item deletes the entire scalar-peaks entity. It requires confirmation.

Change

The Measure/Scalar/Normalize/Change menu item opens a control panel that shows the name of the current scalar entity. Only one scalar entity can be current at a time. To change to another scalar entity, enter a different name and select OK.

Normalize/View

The Measure/Scalar/Normalize/View menu item creates a spreadsheet of the scalar entity. This allows easy viewing of the individual scalar peaks and interatomic distances. While in the spreadsheet, you can edit the distance.


Measure/DISCOVER Restraints

One of the principal goals of analysis of 2D NOESY experiments is accumulating interatomic distance restraints for use in various molecule structure-determination studies. FELIX provides a set of menu items for turning volume buildup rates into distance restraints. Once you have an entity of scalar peaks, you are ready to generate distance bounds from assigned cross peaks. You can generate restraints in DGII/Discover format or in XPlor format, using the Measure/DISCOVER Restraints and Measure/X-PLOR Restraints menu item, respectively.

Several restraint classes are supported in FELIX. The basic 2D NOESY peaks can be used in structure generation and refinement as NOE-distance restraints. Based on an assigned peak entity and measured volumes (optionally buildups), FELIX can create new restraints, interactively show one restraint, and use a list of violations to recalculate restraints.

NOE Distance Define

The NOE Distance Define option creates a new entity (msi:noe_dist) of distance restraints from volume buildup rates. Defining restraints overwrites any existing restraints if Action in the control panel is set to New. If Action is set to Append, the restraints produced are appended to the restraint entity. The entity name containing the scalar information used in calculating the restraints is then input. You then input values for the Lower Force Constant, Upper Force Constant, and Maximum Force. Then you select the method used to derive the buildup rates. You can choose between using a Single mixing time or a "best fit" rate based on the First N mixing times, as calculated by a first-order polynomial (Straight Line), a Second Order Polynomial, or an Empirical Fit. In an Empirical Fit calculation, FELIX uses a scalar entity containing a series of cross-peak intensities that correspond to a range of interatomic distances to determine an empirical relationship between NOE intensity and distance.

Although NOE volumes, in reality, increase exponentially and are damped by the exponential T2 relaxation, there are simply not enough mixing times to yield a robust fit to such an exponential-type function. Since only the initial buildup rate is needed (not the entire function), FELIX provides fit functions that can robustly calculate the initial rate from a small set of volume observations.

You can define, for the case of a symmetric spectrum, which peaks are to be considered in the restraint generation: all peaks or peaks on a specific side of the diagonal (for up to three user-defined regions) or only the lower-intensity peaks (symmetry selection).

All the above methods (except Empirical Fit) use the inverse 6th-power relation between buildup rate and interatomic distance to calculate a single distance in angstroms from the volume buildup rate for each cross peak and the scaling constant derived from the scalar peaks. That one resultant distance is then used to create a restraint having lower and upper distance bounds, based on the chosen method.

The Exact Distance method creates an entity where each restraint is an exact distance. A control panel asks for the minimum and maximum distance allowed for any restraint. For all restraints, the lower- and upper-bound distances are both set to the one calculated distance.

The Strong-Medium-Weak Bins method creates an entity where each restraint is grouped into one of three distance-bound bins. A control panel asks for the minimum and maximum distances allowed, the distance boundary between the strong and medium bins, and the distance boundary between the medium and weak bins. For all restraints, the lower and upper bounds are one of these three explicit distance ranges, depending on which bin the calculated distance falls in.

The Van Der Waals-Exact method creates an entity where each restraint uses a generic van der Waals hard-sphere radius for the lower bound and sets the upper bound to be the calculated distance. A control panel asks for the van der Waals lower bound and the maximum upper bound.

The Percentage of Distance method creates an entity where each restraint uses a percentage of the calculated distance as the lower and upper bounds. A control panel asks for the minimum and maximum distances, a lower-bound percentage, and an upper-bound percentage. For each restraint, the lower bound is obtained by subtracting the lower-bound percentage of the calculated distance from the calculated distance, while the upper bound is obtained by adding the upper-bound percentage of the calculated distance to the calculated distance. In this manner, short distances translate into narrower bounds while longer distances have wider bounds.

Also, you may specify that only the non-overlapped peaks be used in creating restraints. You can choose what percentage of the peak box area overlap is to be handled differently, by setting the Area Threshold. You can then discard those peaks (set Partial Overlap to Discard) or use a different method to turn them into restraints (e.g.: Use as Qual).

NOE Distance Calculate One

The NOE Distance Calculate One menu item is similar to the NOE Distance Define menu item above, except that only a single restraint is calculated and the restraint value is not saved to the database. A cursor is used to specify the cross peak used in the restraints calculation.

NOE Distance Redefine

The Measure/DISCOVER Restraints/NOE Distance Redefine menu item serves as a refinement tool. After a set of peaks is assigned and a set of restraints is extracted based on that assignment, you can try to generate structures by using distance geometry (Insight II/NMR_Refine/DGII) or simulated annealing (Insight II/NMR_Refine/MD_Schedule). After a successful run, several hot-spots are normally discovered, for example, certain assignments or restraints may not be right. If so, the NOE Distance Redefine menu item can help to loosen, tighten, or delete some restraints showing the highest violations (or the highest number of violations within a family). To do this, you must first load the restraints on all the refined molecules into Insight II using the NMR_Refine/Restraints/Read molname* command. Then you may execute the NMR_Refine/Distance/List command. The provided numvioltofelix script redirects the output into another file. Then you can use the Measure/DISCOVER Restraints/NOE Distance Redefine menu item on this file. You should specify the Restraint entity you want to work with (usually the msi:noe_dist) and the Buildup Rate Calculation Method. Then the program shows the violation table, from which you can zoom in on each peak for which the defined restraint was violated. From the table you can see the calculated distance, the restrained values, and the violation statistics. You then can select an Action: Leave as is, Redefine bounds, Delete Restraints.

NOE volume Define

From an assigned peak entity and the corresponding volume entity, FELIX can calculate NOE volume restraints, which contain volume lower bounds and upper bounds as restraining entities. These can then be used in Discover.

The NOE Volume Define menu item calculates and stores NOE-volume restraints for all the assigned peaks.

NOE Distance Overlap Define

In certain instances some peaks can have multiple possible assignments. Those assignments (made in Assign) can be used in Discover to help in the refinement. These restraints are called NOE distance overlapped restraints.

You must have already defined singly assigned peaks as scalar peaks and also have defined the volumes to be measured before you can generate overlapped restraints. You can use the NOE Distance Overlap Define option to define such restraints from 2D NOE spectra or from heteronuclear edited 3D or 4D NOE spectra. Each NOE distance overlap restraint contains a set of possible atom name pairs (multiple possible assignments), as well as an effective distance upper and lower bound. You can export these restraints to the Insight II program and use them as ambiguous restraints in a Discover simulated annealing or rMD, rEM run.

NOE Volume Overlap Define

The peaks with multiple possible assignments can be used in Discover directly - that is, without turning them into effective distances - to help in the refinement. These restraints are called NOE volume overlapped restraints.

You must have already measured the volumes (buildups) before you can generate overlapped restraints. You can use the NOE Volume Overlap Define option to define such restraints from 2D NOE spectra or from heteronuclear edited 3D or 4D NOE spectra. Each NOE volume overlap restraint contains a set of possible atom name pairs (multiple possible assignments), as well as restraining volume(s). You can export these restraints to the Insight II program and use them as ambiguous restraints in a Discover simulated annealing or rMD, rEM run.

3J Dihedral

After measuring 3J couplings using the above described menu items and then assigning peaks, you can create 3J-dihedral restraints using the 3J Dihedral option.


Measure/X-PLOR Restraints

NOE

Several restraint classes are supported in FELIX. The basic 2D NOESY peaks can be used in structure generation and refinement as NOE-distance restraints. Based on an assigned peak entity and measured volumes (optionally buildups), FELIX can create new restraints, interactively show one restraint, and use a list of violations to recalculate restraints.

The Define option creates a new entity (msi:noe_dist) of distance restraints from volume buildup rates.

Note:
Defining restraints overwrites any existing restraints if the Action parameter is set to New.

If Action in the control panel is set to Append, the restraints produced are appended to the restraint entity. The entity name containing the scalar information used in calculating the restraints is then input. Then you select the method used to derive the buildup rates. You can choose between using a Single mixing time or a "best fit" rate based on the First N mixing times, as calculated by a first-order polynomial (Straight Line), a Second Order Polynomial, or an Empirical Fit. In an Empirical Fit calculation, FELIX uses a scalar entity containing a series of cross peak intensities that correspond to a range of inter-atomic distances to determine an empirical relationship between NOE intensity and distance.

Although the NOE volumes, in reality, increase exponentially and are damped by the exponential T2 relaxation, there are simply not enough mixing times to yield a robust fit to such an exponential-type function. Since only the initial buildup rate is needed (not the entire function), FELIX provides fit functions that can robustly calculate the initial rate from a small set of volume observations.

You can define, for the case of a symmetric spectrum, which peaks are to be considered in the restraint generation: all peaks or peaks on a specific side of the diagonal (for up to three user-defined regions) or only the lower-intensity peaks (symmetry selection).

All the above methods (except Empirical Fit) use the inverse 6th-power relation between buildup rate and interatomic distance to calculate a single distance in angstroms from the volume buildup rate for each cross peak and the scaling constant derived from the scalar peaks. That one resultant distance is then used to create a restraint having lower and upper distance bounds, based on the chosen method.

The Exact Distance method creates an entity where each restraint is an exact distance. A control panel asks for the minimum and maximum distances allowed for any restraint. For all restraints, the lower- and upper-bound distances are set to one calculated distance value.

The Strong-Medium-Weak Bins method creates an entity where each restraint is grouped into one of three distance bound bins. A control panel asks for the minimum and maximum distances allowed, the distance boundary between the strong and medium bins, and the distance boundary between the medium and weak bins. For all restraints, the lower and upper bounds are one of these three explicit distance ranges, depending on which bin the calculated distance falls in.

The Van Der Waals-Exact method creates an entity where each restraint uses a generic van der Waals hard-sphere radius for the lower bound and sets the upper bound to be the calculated distance. A control asks prompts for the van der Waals lower bound and the maximum upper bound.

The Percentage of Distance method creates an entity where each restraint uses a percentage of the calculated distance as the lower and upper bounds. A control panel asks for the minimum and maximum distances, a lower-bound percentage, and an upper-bound percentage. For each restraint, the lower bound is obtained by subtracting the lower-bound percentage of the calculated distance from the calculated distance, and the upper bound is obtained by adding the upper-bound percentage of the calculated distance to the calculated distance. In this manner, short distances are translated into narrower bounds, and longer distances have wider bounds.

You may specify that only non-overlapped peaks be used in creating restraints. You can choose what percentage of the peak box area overlap is to be handled differently by setting the Area Threshold parameter. You can then discard those peaks (set Partial Overlap to Discard) or use a different method to turn them into restraints (for example, Use as Qual).

Ambiguous NOE

In certain instances some peaks can have multiple possible assignments. Those assignments (made in Assign) can be used in XPlor to help in the refinement. These restraints are called ambiguous NOE restraints.

You must have already defined singly assigned peaks as scalar peaks and also have defined the volumes to be measured before you can generate overlapped restraints. You can use the Ambiguous NOE and Define options to define such restraints from 2D NOE spectra or from heteronuclear edited 3D or 4D NOE spectra. Each ambiguous NOE restraint contains a set of possible atom name pairs (multiple possible assignments), as well as effective distance upper and lower bounds. You can export these restraints to the Insight II program and use them as ambiguous restraints in a XPlor simulated annealing or rMD, rEM run.

Dihedral

After measuring 3J couplings using the above described menu items and assigning peaks, you can create dihedral restraints using the Dihedral option.

NOE-Intensity

Using an assigned peak entity and the corresponding volume entity, FELIX can calculate NOE intensity restraints, which contain volume lower and upper bounds as restraining entities in the NOE-Intensity and Define options. These can then be used in XPlor.

Ambiguous NOE-Intensity

Peaks with multiple possible assignments can be used in XPlor directly (that is, without turning them into effective distances) to help in the refinement. These restraints are called ambiguous NOE-intensity restraints.

You must have already measured the volumes (buildups) before you can generate overlapped restraints. You can use the Ambiguous NOE-Intensity and Define options to define such restraints from 2D NOE spectra or from heteronuclear edited 3D or 4D NOE spectra. Each ambiguous NOE intensity restraint contains a set of possible atom name pairs (multiple possible assignments), as well as restraining volume(s) bound. You can export these restraints to the Insight II program and use them as ambiguous restraints in an XPlor simulated annealing or rMD, rEM run.

NOE-NOE

If you have an assigned 3D NOE-NOE spectrum where you measured volumes, you can turn them into 3D NOE-NOE restraints using the NOE-NOE and Define options.


Help pulldown

The Help pulldown (<Alt>-h) contains menu items that provide information on using the FELIX program.


Help/About

The Help/About menu item (<Alt>-ha) gives you information about the current version of FELIX.


Help/Topic

The Help/Topic menu item (<Alt>-ht) opens a Netscape browser window and allows you to access the online documentation.


Help/Keypad

The Help/Keypad menu item (<Alt>-hk) gives you information about the currently available navigation menu items and their keypad shortcuts.


Help/Command/Edit

The Help/Command/Edit menu item (<Alt>-hc) gives you information about switching between command-line mode and shortcut-keypad mode in the main FELIX window, as well as information about switching between navigation and edit mode in the spreadsheet viewer windows.