Peak Table¶

The peak table features in many parts of the workflow, and contains (with a few exceptions) the following columns.

Peak table fields¶
Abbreviation	Description
h	h Miller index
k	k Miller index
l	l Miller index
x	x coordinate of unmerged peak (pixels)
y	y coordinate of unmerged peak (pixels)
frame	frame coordinate of unmerged peak
I (sum)	Integrated (corrected) intensity from pixel sum integration
I (profile)	Integrated (corrected) intensity from profile integration
\(\sigma\) (sum)	Variance from pixel sum integration
\(\sigma\) (profile)	Variance from profile integration
\(I/\sigma\) (sum)	Strength from pixel sum integration
\(I/\sigma\) (profile)	Strength from profile integration
B (sum)	Integrated backround from pixel sum integration
B (profile)	Integrated backround from profile integration
\(\sigma\) (B, sum)	Background variance from pixel sum integration
\(\sigma\) (B, profile)	Background variance from profile integration
Gradient	Background gradient (image plane only, pixel sum integration only)
\(\sigma\) (Grad.)	Background gradient variance
Data set	Data set associated with this peak
Unit cell	Unit cell associated with this peak
d	Resolution associated with this peak (d from Bragg’s law)
Reason for rejection	Reason, if any, why this peak is marked as invalid
Caught by filter	Whether this peak was caught by a filter

Note that not all of these columns appear in every instance of the peak table, but this is an exhaustive list of possible fields.

Rejection flags¶

A peak flagged as “invalid” will not be integrated and is therefore not included in the merge statistics. The data quality statistics (R-factors and CC-values) can be indefinitely improved by simply regjected peaks that make them worse, so some measure of how many peaks were rejected is necessary to ensure the integrity of the data; this is the completeness, and it is the number of valid peaks as a fraction of the total number of possible peaks. There are many possibly reasons why a peak might be rejected; these are listed below.

Rejected during peak finding¶

Masked by user - the peak intersects a mask applied to the detector image. Such masks are applied to exclude problematic regions of the detector, such as seams between detector plates and the area surrounding the beam stop.
Too many or few detector counts - a blob must have a number of counts within a given range to qualify as a valid peaks. The upper limit is used to exlude pixels that have been saturated, i.e. the count is beyond the limit of the detector.
Peak centre outside frame range - the centre of mass of the blob falls outside the angular region encompassed by the detector images.
Peak centre outside detector image - the centre of mass of the blob falls outside the bounds of the detector image.

Rejected during integration¶

Integration failed - This is a fall-through condition met if the integration fails but not for any other of the listed reasons. Should not generally happen.
Too few points to integrate - the integration region contains too few counts to reliably integrate.
No unit cell assigned - the given peak has no unit cell assigned and can not be handled.
No associated data set - The given peak has no data set assigned and can not be handled.
Integration region extends beyond image/frame range - The integration region extends either beyond the detector image, or outside the frame (rotation) range.
Negative, zero, or undefined background sigma - certain rare circumstances might lead to a mathematically undefined sigma.
Peak contains saturated pixel - A 16-bit detector can hold a maximum of 65535 counts per pixel. Moreover, there is physical limit to the number of counts a detector pixel can reach. If the integrator is instructed to reject peaks containing such pixels, this reason is used.
Adjacent peak intensity region overlaps this peak - if two peak intensity regions overlap, both peaks are rejected.
Centre of mass of peak is inconsistent - the centre of mass of the peak is mathematically undefined.
Covariance matrix of peak is inconsistent - the shape of the ellipsoid defining the peak, i.e. the covariance matrix, is mathematically undefined.
Shape of peak is too small or large - the eigenvalues of the covariance matrix of the ellipsoid are too small or large.
Failed to find a minimum of I/Sigma - specific to the \(\sigma/I\) integrator, which requires a minimum in \(I/\sigma\) to determine the extents of the integration region.
Peak centre moved beyond bounds of data set - specific to the Gaussian integrator: the peak centre may move outside the bounds of the image or frame range during integration.
Pearson coefficient of fit is too low - specific to the Gaussian Integrator: the fit of the peak shape is not good enough.
No shape model found - No shape model was specified during profile integration.

Rejected during shape prediction¶

No neighbouring profile to compute shape - There are no neighbouring peaks within the specified pixel and frame range cutoffs with which to compute a mean covariance.
Too few neighbouring profiles to compute shape There are too few neighbouring peaks within the specified pixel and frame range cutoffs with which to compute a mean covariance.

Other¶

Frame coordinate interpolation failed - The fractional frame value (i.e. specific rotation angle) cannot be determined because the ellipsoid extends beyond the frame range.
Manually unselected by user - peak can be manually deselected form the detector image by selecting one or many peaks with the rectangle select tool, then pressing backspace.
Outside indexing tolerance - The non-integer Miller indices (the product of a reciprocal vector and reciprocal basis) are too far from the specified integer Miller indices.