Density values after alignment not seem correct

5 posts / 0 new
Last post
hzhan3
Density values after alignment not seem correct

Hi all, 

Sorry for the stupid questions...after I used unblur from cisTEM to align the frames for each tomography tilt, the density values seem too small between -2 to 2 with Mean density at 0.94703 e-9. Briefly, I applied gain reference while I imported each set of frames, and bin 2 as suggested, other value in expert mode I kept as default. I am wondering: 1) whether the scaling is not correct? 2) How can I check scaling? 3) Is there a way to change the output mode to integer?

Thanks!!!

Hong

timgrant
Hi Hong,

Hi Hong,

This is possibly just because the statistics in the MRC file are not written out by cisTEM.  cisTEM ignores all statistics, and so is not very good at writing them.

If you use the command line program "reset_mrc_header" that is included in the cisTEM bin directory on one of these files - does it fix the problem?

Thanks,

Tim

hzhan3
Hi Tim,

Hi Tim,

I tried reset_mrc_header in the commend line, it only allows me to do New pixel size? it also changed the origin on x,y,z and the density still the same. Is the negative density value normal? I think it was calculated in the real space?

RO image file on unit   1 : 01.mrc     Size=      55622 K

 Number of columns, rows, sections .....    3838    3710       1
 Map mode ..............................    2   (32-bit real)              
 Start cols, rows, sects, grid x,y,z ...    0     0     0    3838   3710      1
 Pixel spacing (Angstroms)..............   2.600      2.600      2.600    
 Cell angles ...........................   90.000   90.000   90.000
 Fast, medium, slow axes ...............    X    Y    Z
 Origin on x,y,z .......................    0.000       0.000       0.000    
 Minimum density .......................  -1.0396    
 Maximum density .......................   2.0160    
 Mean density ..........................  0.94703E-09
 RMS deviation from mean................  0.32214    
 tilt angles (original,current) ........   0.0   0.0   0.0   0.0   0.0   0.0
 Space group,# extra bytes,idtype,lens .        1        0        0        0

     1 Titles :
** GuiX **                                                                     

hong@hongzhan:~/Documents/test/testunblur$ reset_mrc_header

        **   Welcome to ResetMRCHeader   **

                Version : 1.00
               Compiled : Dec  2 2017
                   Mode : Interactive

Input image file name [input.mrc]                  : 01.mrc
New pixel size [1.0]                               : 2.6
(18h:12m10s)                

All done.
hong@hongzhan:~/Documents/test/testunblur$ ls
01.mrc
hong@hongzhan:~/Documents/test/testunblur$ header 01.mrc

 RO image file on unit   1 : 01.mrc     Size=      55622 K

 Number of columns, rows, sections .....    3838    3710       1
 Map mode ..............................    2   (32-bit real)              
 Start cols, rows, sects, grid x,y,z ...    0     0     0    3838   3710      1
 Pixel spacing (Angstroms)..............   2.600      2.600      2.600    
 Cell angles ...........................   90.000   90.000   90.000
 Fast, medium, slow axes ...............    X    Y    Z
 Origin on x,y,z .......................   -4989.      -4823.       0.000    
 Minimum density .......................  -1.0396    
 Maximum density .......................   2.0160    
 Mean density ..........................  0.94703E-09
 RMS deviation from mean................  0.32214    
 tilt angles (original,current) ........   0.0   0.0   0.0   0.0   0.0   0.0
 Space group,# extra bytes,idtype,lens .        1        0        0        0

     1 Titles :
** GuiX **                                

Thanks!

Hong

timgrant
Hi Hong,

Hi Hong,

That should means those density values are corrected.  Is there some reason you believe them to be incorrect?  Negative density values are fine if the image has been zero floated (set so the average value is 0).  As the addition / subtraction of a constant in real space changes only the central pixel in Fourier space, it basically changes nothing.

Tim

hzhan3
Hi Tim, 

Hi Tim, 

Thanks for the explanation! I will proceed with the processing in unblur. 

Thanks!

Hong

Log in or register to post comments