pipeline:isac:sxcompute_isac_avg

This version (2019/04/02 10:39) is a draft.
Approvals: 0/1

Beautifier: Perform local 2D alignment of ISAC2 2D clustering results using the original pixel size and full CTF correction.


Usage in command line

sp_compute_isac_avg.py  --stack=FILE_PATH  --isac_dir=DIRECTORY_PATH  --output_dir=DIRECTORY_PATH  --pixel_size=PIXEL_SIZE  --radius=PARTICLE_PIXEL_RADIUS  --pw_adjustment=METHOD_OR_FILE_PATH  --fl=CUTOFF_FREQ_METHOD  --noctf  --local_alignment  --B_start=LOWER_RESOLUTION_BOUND  --Bfactor=B_FACTOR  --xr=PIXEL_RANGE  --ts=PIXEL_STEP  --fh=HIGH_FREQ_SERACH_LIMIT  --maxit=MAX_ITERS  --navg=NUM_OF_AVERAGES


sp_compute_isac_avg.py exists only in MPI version.

There are four methods to adjust the power spectrum of resultant average images, as specified by option pw_adjustment.


1. Adjust to analytical model (default method):

mpirun -np 48 sp_compute_isac_avg.py --stack=bdb:data --isac_dir=outdir_isac --output_dir=outdir_beautify2d --radius=120


2. Adjust to B-factor:

mpirun -np 48 sp_compute_isac_avg.py --stack=bdb:data --isac_dir=outdir_isac --output_dir=outdir_beautify2d --radius=120 --pw_adjustment=bfactor --Bfactor=40.0 --fl=0.2


2. Adjust to rotational average:

mpirun -np 48 sp_compute_isac_avg.py --stack=bdb:data --isac_dir=outdir_isac --output_dir=outdir_beautify2d --radius=120 --pw_adjustment=rot_avg_1d_pw_spectrum.txt


4. No adjustment:

mpirun -np 48 sp_compute_isac_avg.py --stack=bdb:data --isac_dir=outdir_isac --output_dir=outdir_beautify2d --radius=120 --pw_adjustment=no_adjustment


Main Parameters

--stack
Original image stack: Data stack that used for the associated ISAC2 run. The particle images in this stack are used to create the full-sized class averages. (default required string)
--isac_dir
ISAC2 run directory: Path to the output directory of the associated ISAC2 run. This is input directory for this command. (default required string)
--output_dir
Output directory: The automatically-created output directory will contain results. By default, the program uses sharpen_DATA_AND_TIME for the name. If this is the same as ISAC2 run directory, the program automatically creates sharpen subdirectory under the ISAC2 run directory. (default none)
--pixel_size
Pixel size [A]: The pixel size of particle images in input particle stack for the associated ISAC2 run. Use 1.0 in case of negative stain data. (default -1.0)
--radius
Particle radius [Pixels]: There is no default radius. (default -1)
--noctf
CTF correction: Indicate if full CTF correction should be applied or not. Always use the CTF correction for cryo data, but not for negative stained data. By default, do full CTF correction. (default False question reversed in GUI)


Advanced Parameters

--local_alignment
Local alignment: Indicate if local alignment should be applied or not. (default False value reversed in GUI)
--fl
Low-pass filter frequency [1/Pixel]: Cutoff frequency of low-pass filter. =-1.0, do not apply the low-pass filter; =0.0, apply low pass filter to initial ISAC2 resolution; =1.0, to resolution after local alignment; else use user-provided cutoff in absolute frequency (>0.0 and ⇐0.45). (default -1.0)
--pw_adjustment
Power spectrum adjustment method: Specify the method for the power spectrum (PWS) adjustment of 2-D averages to enhance averages. ='analytical_model' adjusts PWS to an analytic model; ='bfactor' adjusts PWS using B-factor; ='FILE_PATH' adjusts PWS to rotationally averaged 1D power spectrum stored in the text file; ='no_adjustment' skips adjustment. (default analytical_model)
--B_start
Lower bound for B-factor estimation [A]: Lower resolution bound of power spectrum for B-factor estimation. Specific to adjust to B-factor method. (default 45.0)
--Bfactor
Use ad-hoc B-factor [A^2]: Skip the automatic estimation and use user-provided ad-hoc B-factor (e.g. 25.0[A^2]) for the enhancement. By default, the program automatically estimates B-factor. Specific to adjust to B-factor method. (default -1.0)
--xr
Local search range [Pixels]: Translational search range for local alignment. (default -1.0)
--local_alignment==True
--ts
Local search step [Pixels]: Translational search step for local alignment. (default 1.0)
--local_alignment==True
--fh
High frequency search limit [1/Pixel]: High frequency search limit in absolute frequency for local alignment. (default -1.0)
--local_alignment==True
--maxit
Local alignment iterations: The number of iterations for local aligment. (default 5)
--local_alignment==True
--navg
Number of averages: The number of averages to be process, starting from the first image. By default, uses all ISAC2 average images. (default 1000000)


Important Output Files

File Name Discription
class_averages.hdf Computed averages in original image size. Each average has attributes “members” and “n_objects” in the image header.
ordered_class_averages.hdf Sorted version of class_averages.hdf. Averaged images are sorted by their similarity and re-arranged in a cyclic manner.
params_avg_*.txt 2D alignment parameters computed by ISAC2.
init_isac_params.txt 2D alignment parameters of all particles computed by ISAC2.
ali2d_local_params_avg_*.txt 2D alignment parameters computed by local alignment of this program.
ali2d_local_params*.txt 2D alignment parameters of all particles computed by local alignment of this program.
FH_list.txt Initial ISAC2 resolution and resolution after local alignment.




Not published yet.


2018/07/10 Toshio Moriya

Reflected the interface changes.

- Added Options

--pw_adjustment
Power spectrum adjustment method: Specify the method for the power spectrum (PWS) adjustment of 2-D averages to enhance averages. ='analytical_model' adjusts PWS to an analytic model; ='bfactor' adjusts PWS using B-factor; ='FILE_PATH' adjusts PWS to rotationally averaged 1D power spectrum stored in the text file; ='no_adjustment' skips adjustment. (default analytical_model)

- Removed Options

--B_enhance
Adjust to B-factor: Adjust the power spectrum of resultant average images using B-factor to enhance averages. Specific to adjust to B-factor method. (default False)
--adjust_to_given_pw2
Adjust to rotational average: Adjust the power spectrum of resultant average images to the user-provided 1-D reference power spectrum. Specific to adjust to rotational average method. (default False)
--adjust_to_analytic_model
Adjust to analytical model: Adjust the power spectrum of resultant average images to an analytical model. Specific to adjust to analytical model. (default False)
--no_adjustment
No adjustment: Do not adjust the power spectrum of resultant average images. Use this option to skip all power spectrum adjustment methods and simply compute class averages with the original pixel size. Specific to no adjustment method. (default False)


Zhong Huang


Category 1:: APPLICATIONS


sparx/bin/sp_compute_isac_avg.py


Beta:: Under evaluation and testing. Please let us know if there are any bugs.


There are no known bugs so far.


  • pipeline/isac/sxcompute_isac_avg.txt
  • Last modified: 2019/04/02 10:39
  • by lusnig