Content-type: text/html Manpage of calcFSC


Section: User Commands (1)
Index Return to Main Contents


calcFSC - calculate the Fourier Shell Correlation of an alignment  


calcFSC file.prm
calcFSC file.prm repCount
calcFSC file.prm repCount flgShuffle  


calcFSC calculates the Fourier Shell Correlation (FSC) resulting from the particle alignment specified in parameter file 'file.prm'. Particles to be processed are selected using an approach similar to that used by averageAll. The primary difference is that the number of particles to process is specified by parameter "nParticles", rather than "lstThresholds". Particles selected for processing are divided into 2 equal-sized groups after alignment, and the FSC computed by comparing their averages in Fourier space. Result are stored in text files and can be graphed using plotFSC. Like 'lstThresholds', 'nParticles' can be a vector, in which case an FSC curve will be calculated for each entry.

Computing an FSC in this fashion (i.e. based on a common, mutual alignment) can lead to bias and overly optimistic results. For a more reliable estimate, consider manually splitting the data prior to alignment and then using calcUnbiasedFSC(1).

is the name of the PEET parameter file.
is an optional positive integer specifying the number of times the FSC calculation should be repeated. If 1 or greater, time-based initialization of the pseudo-random number generator will be used, so subsequent runs will give different results. If greater than 1, the caculated FSC curve will be the average of 'repCount' runs with different pseudo-random assortment for each run. If greater than 2, the standard deviation of the mean estimated FSC will also be computed and saved to file arrFSCC.txt, for use by plotFSC.
If flgShuffle is non-zero, return different results one each invocation, even if repCount is 1. (Default = 0). The parameter file is a MATLAB readable text file that specifies the parameter values in MATLAB syntax. See the PEET man page for a description of the parameter file options and syntax. calcFSC uses the 'flgNormalize' , 'fnVolume', 'fnModParticle', 'fnOutput', 'dPhi', 'lstThresholds', 'szVol' , 'tiltRange', 'edgeShift', 'flgRemoveDuplicates', flgRandomize, flgVolNamesAreTemplates, and 'debugLevel' parameters described there. In addition, calcFSC supports the 'includeList', 'excludeList', and 'selectClassID' parameters described on the averageAll man page. It also supports additional, optional parameters specific to calcFSC listed below. These optional parameters, some of which are shared with calcSSNR, all have reasonable default values. Extra parameters used to control the other programs are ignored by calcFSC and vice versa, so a single parameter file should suffice even when modifying some parameters. Sample parameter files are available in $PARTICLE_DIR/templates.

calcFSC needs to be run in the same directory where alignSubset and averageAll were run.

nParticles = < int | vector >
The number of particles in each group; a vector indicates that multiple results are requested. If there are n particles specified by the alignment structure, the maximum value in nParticles should be floor(n/2). The default is an arithmetic sequence of 5 integers that start from 1/10 of the last entry in the 'lstThresholds' array described in PEET man page and go up to floor(n/2).
freqShells = vector
The upper limit of each frequency shell. The default is [0.025:0.025:0.5].
windowName = string
The type of the mask window to apply to the data, 'gaussian', 'hamming', etc. The default is 'gaussian'. If windowName is 'file', windowParams must be a string with the name of an MRC file containing the mask. A windowName of 'none' can be used to disable masking.
windowParams = vector
The parameters to pass to the window generation function. For a gaussian window the parameters are [radius sigma]; see gaussianMask in MATLAB help. The default is [0.35 0.05].
windowCenter = vector
The center of the mask window (default: [ ] = volume center). The default is floor(szVol/2)+1.
fnHalfVolume = string
The base name of the half volumes to write out (default: '', don't write out the half volumes).
fscMaskOrigin = vector
Specifies the origin of your FSC mask in the average volume coordinate system that has the X-Y origin at lower left corner and Z origin at slice 0. The center of the average volume will have coordinates szVol/2.
fscMaskSize = vector
Specifies the FSC mask size in the X, Y, and Z dimension. Only voxels inside this rectangular mask are included in the FSC calculation.


Sample parameter files are available in $PARTICLE_DIR/templates.

calcFSC stores results in 4 or 5text files in the current working directory. 'szVoxel.txt' contains the pixel size in Angstroms. 'nParticles.txt' contains the the number of particles used, which may differ from those requested in parameter 'nParticles'. 'freqShells.txt' has the value for parameter 'freqShells'. 'arrFSCC.txt' contains a 2 dimensional array of correlation scores. Each of its columns corresponds to a number of particles used, with rows corresponding to frequencies in 'freqShells'. If repCount is greater than 1, pointwise standard deviations of the mean will be stored in an optional 'arrSigma.txt' file.  


Rick Gaudette
Quanren Xiong
John Heumann  


PEET(1), plotFSC(1), prmParser(1), prepareEM(1), prepareRef(1), mergeEM(1), alignSubset(1), averageAll(1), calcUnbiasedFSC(1), and calcSSNR(1)




This document was created by man2html, using the manual pages.
Time: 21:10:05 GMT, March 20, 2017