Dealing with K2 data from SerialEM
A few niggles that need to be dealt with when importing serialEM K2 data. Last updated: February 2020.
We typically use serialEM for K2 data collections in part due to its ability to write out movies in a compressed tif format. Considering the rapid write out speeds that we can achieve now with K2 and beam tilt corrections, minimising movie size is crucial for a healthy storage space.
Data Transfer
This will be done automatically in future. For the time being, it is essential that you make folders adhering to our new naming conventions (see data transfer tab or top level for strubiEM pages for a descriptor), and email emsupport@strubi.ox.ac.uk asap. There is limited space on the K2 pc and thus transfer must be done on the fly and shortly after collection set up in most cases.
Formatting inputs
For this section, imod is your friend. Please check that you module load this initially. For the defects file, the latest version of imod is currently only on Strubi and thus you will need to transfer the defect file to a strubi computer (e.g. coke) first to do the conversion.
Defects
For processing in relion, you will need to convert the defects file into a map-format. This can be achieved with the latest version of imod (4.10).
Gain reference
The gain reference is written out as a dm4 file. Convert this to an mrc file using imod.
dm2mrc serialEMgainref.dm4 gain.mrc
Importing Movies in cryosparc and relion
Now you are good to go to relion of cryosparc for processing.
Note that, to apply the gain ref, it is necessary to apply a 90 degree rotation and flip in y (or 270 degree rotation and flip in x) for cryosparc. Note that the rotation value in cryosparc is expressed as a number, where 1 represents 90 deg i.e. for a 90 deg rotation, you need to put '1', and for 270 you put a '3'. In relion, you only need to apply a 270 degree rotation.
It is good practice to check the header of one of your serialEM movies to establish what transformation is necessary. This can be established by checking the 'r/f' value in the 'header' command (imod).
module load imod/4.9.9_RHEL7-64_CUDA8.0 #If you have not done this already.
header example.tif
RO image file on unit 1 : /data/bowmore/krios/stuart/Frames/example.tif Size= 124120 K
This is a TIFF file.
Number of columns, rows, sections ..... 3710 3838 30
Map mode .............................. 6 (unsigned 16-bit integer)
Start cols, rows, sects, grid x,y,z ... 0 0 0 3710 3838 30
Pixel spacing (Angstroms).............. 0.8873 0.8873 0.8873
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 ....................... 0.0000
Maximum density ....................... 7.0000
Mean density .......................... 3.5000
tilt angles (original,current) ........ 0.0 0.0 0.0 0.0 0.0 0.0
Space group,# extra bytes,idtype,lens . 0 0 0 0
3 Titles :
SerialEMCCD: Dose frac. image, scaled by 1.00 r/f 3
example_00000.dm4
defects_example_00000.txt
No rotation/noflip | 0 |
90 deg rotation/noflip | 1 |
180 deg rotation/noflip | 2 |
270 deg rotation/noflip | 3 |
No rotation/Flip around Y | 4 |
90 deg rot/flip around Y | 5 |
180 deg rot/Flip around Y | 6 |
270 deg rot/Flip around Y | 7 |
For cryosparc collections, we have found that with recent data collections in Strubi (February 2020 i.e. when the 'hair' was present), it is necessary to apply a flip around Y and 90 deg rotation, or flip around X and 270 deg rotation, despite the movie header suggesting otherwise. For cryosparc, the rotation option is a discrete number from 1-4 (i.e. 1 for 90 deg , 3 for 270 deg).