Main Panel¶
Clone: Clone popup window
Help: Show popup help document
Close: Close popup
** Peak Settings and Non-Interactive Peak Finding **
The purpose of this dialog is to allow the user to select settings for finding and integrating peaks, and also to be able to find peaks in an arbitrary region that is specified in a table rather than via a spectrum window.
** Find Parameters tab **
This can be used to specify how peak finding works.
First of all, you can search for just positive peaks, just negative peaks or both, and the default is that it is just positive peaks. However, this is further filtered by what the contour levels are. If there are no positive contour levels for a given spectrum then positive peaks are not found even if this dialog says they can be, and similarly if there are no negative contour levels for a given spectrum then negative peaks are not found even if this dialog says they can be.
The peak finding algorithm looks for local extrema (maximum for positive peaks and minima for negative peaks). But on a grid there are various ways to define what you mean by an extremum. Suppose you are trying to determine if point p is a maximum (similar considerations apply for minimum). You would want the intensity at all nearby points to be less than or equal to the intensity at p. You can just check points that are just +- one point from p in each dimension, or you can also check “diagonal” points. For example, if you are looking at point p = (x, y) in 2D, then the former would mean checking the four points (x-1, y), (x+1, y) (x, y-1) and (x, y+1), whereas for the latter you would also have to check (x-1, y-1), (x-1, y+1), (x+1, y-1) and (x+1, y+1). In N dimensions the “diagonal” method involves checking 3^N-1 points whereas the “non-diagonal” method involves checking only 2N points. In general the “non-diagonal” method is probably the one to use, and it is the default.
Peaks are only found above (for positive peaks) or below (for negative peaks) some threshold. By default this is determined by the contour level for the spectrum. For positive peaks the threshold is the minimum positive contour level, and for negative peaks the threshold is the maximum negative contour level. However these levels can be scaled up (or down) using the “Scale relative to contour levels” option (default value 1). For example, if you have drawn the contour levels low to show a bit of noise, but do not want the noise picked as peaks, then you could select a scale of 2 (or whatever) to increase the threshold.
The “Exclusion buffer around peaks” is so that in crowded regions you do not get too many peaks near one location. By default the exclusion buffer is 1 point in each dimension, but this can be increased to make the algorithm find fewer peaks.
By default the peak finding only looks at the orthogonal region that is displayed in the given window where peak finding is taking place. Sometimes it looks like a peak should be found because in x, y you can see an extremum, but unless it is also an extremum in the orthogonal dimensions it is not picked. You can widen out the points being examined in the orthogonal dimensions by using the “Extra thickness in orthogonal dims” option, which is specified in points.
The “Minimum drop factor” is by what factor the intensity needs to drop from its extreme value for there to be considered to be a peak. This could help remove sinc wiggle peaks, for example. The default is that the drop factor is 0, which in effect means that there is no condition.
The “Volume method” is what is used to estimate the volume of peaks that are found. The default is “box sum”, which just looks at a fixed size box around the peak centre and sums the intensities in that. The size of the box is set in the table in the Spectrum Widths tab. The “truncated box sum” is the same as “box sum” except that the summing stops in a given direction when (if) the intensities start increasing. The “parabolic” fit fits a quadratic equation in each dimension to the intensity at the peak centre and ad +- 1 points and then uses the equivalent Gaussian fit to estimate the volume.
** Spectrum Widths **
This can be used to specify minimum linewidths (in Hz) for there to be considered a peak to exist in the peak finding algorithm. It is also where the Boxwidth for each dimension in each spectrum is specified.
** Diagonal Exclusions **
This can be used to exclude peaks from being found in regions near the diagonal (so in homonuclear experiments). The exclusion region is specified in ppm and is independent of spectrum.
** Region Peak Find **
This can be used to find peaks non-interactively (so not having to control shift drag inside a spectrum window). The region being analysed is specified in the table. There are two types of conditions that can be specified, “include” for regions that should be included and “exclude” for regions that should be excluded. The regions are specified in ppm.
The “Whole Region” button will set the selected row in the table to be the entire fundamental region of the spectrum.
The “Add Region” button adds an extra row to the table, and the “Delete Region” button removes the selected row.
The “Adjust Params” button goes to the Find Parameters tab.
The “Find Peaks!” button does the peak finding.
Documentation missing
*None*: Sets whether peak picking within spectra find intensity maxima, minima or both maxima and minima
positive and negative: Documentation missing
positive only: Documentation missing
negative only: Documentation missing
*None*: Sets how permissive the peak picking in when searching for intensity extrema; by adding extra points to the selected search region
+-1 in at most one dim: Documentation missing
+-1 allowed in any dim: Documentation missing
Scale relative to contour levels: Threshold above which peaks are picked, relative to the lowest displayed contour; 1.0 means picking exactly what is visible
Exclusion buffer around peaks (in points): The size of the no-pick region, in data points, around existing picked peaks; eliminates duplicate picking
Extra thickness in orthogonal dims (in points): Sets whether to consider any additional planes (Z dimension) when calculating peak volume integrals
Minimum drop factor (0.0-1.0): Documentation missing
Volume method: Selects which method to use to calculate peak volume integrals when peaks are picked; box sizes are specified in “Spectrum Widths”
Documentation missing
Spectrum: The spectrum which determines the widths being shown
Table 1 | |
Dimension | The number of the spectrum dimension to which the settings apply |
Isotope | The nuclear isotope measures in the spectrum dimension |
Minimum Linewidth (Hz) | The smallest value for the linewidth of a peak for it to be picked (Editable) |
Boxwidth | The size of the spectrum region to perform the volume integral over (Editable) |
Documentation missing
Table 2 | |
Isotope | The isotope as measures on the axis of a spectrum window |
Diagonal Exclusion (ppm) | The distance from the homonuclear diagonal line within which no peak picking can occur (Editable) |
Documentation missing
Peak List: Selects which peak list to perform region-wide peak picking for
Table 3 | |
Condition | Documentation missing |
Whole Region: Sets the currently selected region row to cover the whole spectrum
Add Region: Add a new region row, which may them be set for exclusion or inclusion when peak picking large areas
Delete Region: Remove the selected region specification
Adjust Params: Go to the panel for setting the parameters that control how peaks extrema are picked
Find Peaks!: Using the stated regions and parameters, perform region-wide peak picking