Generate RELION Euler angles and rotation origins from UCSF Chimera sessions

Overview

Automatic subtomogram aligment for particle reconstruction not always produces the desired outcome. In order to guide particle reconstruction, the use of a reference may be necessary. By manually superimposing a reference density on the desired location of a subtomogram one can guide alignment, this can be done in visualization software, such as UCSF Chimera. We provide two python scripts that together take the reference alignment information from a UCSF Chimera session and produce a RELION *.star file with initialized Euler angles and rotations origins that transform particles to the reference. This results in the alignment and superimposition of all user-defined subtomogram locations.

Installation

The easiest way to manage the libraries needed to run these scripts is to install Anaconda, and use it to generate the follwing environments:

pychimera, with the libraries necessary to interpret UCSF Chimera session files.

py37, with a version scipy that has the Rotation object.

1. Install Anaconda following the instruction in their website.

2. Clone this repository:

git clone https://github.com/GrotjahnLab/guided_tomo_align.git

cd guided_tomo_align

3. For the pychimera library to work, you need to install UCSF Chimera and direct pychimera to its directory. See pychimeras' help for more information, specially points 2 and 4 in this page. Avoiding launching of the UCSF Chimera GUI may be desirable, specially when working from a computing cluster. For this, make sure you obtain the UCSF Chimera "headless" version. In order to execute the downloaded *.bin file, make it into an executable by running chmod +x ucsfchimera_download.bin. Finally, set the CHIMERADIR environment variable (in bash): export CHIMERADIR=/your/chimera/intallation/dir/

4. Create pychimera and py37 using the provided *.yml files:

conda env create -f ./pychimera.yml

conda env create -f ./py37.yml

Usage of Chimera-to-Euler scripts

1. First, activate the pychimera environment using conda activate pychimera to run chim_session_to_mtx.py. You can print a help message typing:

python ./chim_session_to_mtx.py --help

It will generate the following output:

usage: chim_session_to_mtx.py [-h] --sessions_path SESSIONS_PATH
                          --reference_name REFERENCE_NAME
                          [--particle_name_prefix PARTICLE_NAME_PREFIX]
                          [--particle_name_postfix PARTICLE_NAME_POSTFIX]
                          [--particle_origin PARTICLE_ORIGIN]
                          [--reference_origin REFERENCE_ORIGIN]
                          --particle_angpix PARTICLE_ANGPIX
                          --reference_angpix REFERENCE_ANGPIX --r_box
                          R_BOX --p_box P_BOX [--output_name OUTPUT_NAME]

Use UCSF Chimera *.py sessions to extract the rotation matrix that transforms
particles to references. It is recommended particle densities are named as
they appear in the RELION *.star file. Access to the session MRC files is not
required.

optional arguments:
   -h, --help            show this help message and exit
   --sessions_path SESSIONS_PATH
                    Absolute path to directory where all sessions are
                    stored, e.g., "/User/sessions/"
   --reference_name REFERENCE_NAME
                    Name of reference density in the Chimera sessions as
                    it appears in it. E.g., "low_passed_ribosome.mrc"
   --particle_name_prefix PARTICLE_NAME_PREFIX
                    Prefix used to parse particle densities from Chimera
                    sessions. If all your particles are named
                    "tomogram_<something>.mrc", then use "tomogram_" as
                    prefix to find them in each session.
   --particle_name_postfix PARTICLE_NAME_POSTFIX
                    Postfix used to parse particle densities from Chimera
                    sessions. If all your particles are named
                    "tomogram_<something>_foo.mrc", then use "_foo.mrc" as
                    postfix to find them in each session.
   --particle_origin PARTICLE_ORIGIN
                    Index origin of particle *.mrc if different from
                    0,0,0. It is assumed to be the same for all particles.
                    Enter as X,Y,Z
   --reference_origin REFERENCE_ORIGIN
                    Index origin of reference *.mrc if different from
                    0,0,0. Enter as X,Y,Z
   --particle_angpix PARTICLE_ANGPIX
                    Pixel size, in Angstroms, of particle *.mrc. It is
                    assumed to be the same for all particles.
   --reference_angpix REFERENCE_ANGPIX
                    Pixel size, in Angstroms, of reference *.mrc.
   --r_box R_BOX         Reference box size
   --p_box P_BOX         Particle box size
   --output_name OUTPUT_NAME
                    Name for output file containing a transformation
                    matrix for each particle.

Here's a commandline example:

python ./chim_session_to_mtx.py --sessions_path /path/to/chimerasessions/ --reference_name reference.mrc --particle_name_prefix deconv_lam2_TS4 --particle_angpix 10.88 --reference_angpix 10.88 --r_box 44,44,44 --p_box 44,44,44

NOTE: There is a known issue with some UCSF Chimera installations where the libgfxinfo.so library will produce a "undefined symbol" error. We have been able to fix this by removing the pcre package from the pychimera environment: conda remove --force pcre

2. Deactivate the pychimera environment by typing conda deactivate. Then, activate the py37 environment using conda activate py37 to run mtx_to_star.py.

python ./mtx_to_star.py --help

It will generate the following output:

   usage: mtx_to_star.py [-h] --template_star TEMPLATE_STAR --chimera_mtx_output
                  CHIMERA_MTX_OUTPUT
                  [--rln_particle_particle_csv RLN_PARTICLE_PARTICLE_CSV]
                  [--angpix ANGPIX] [--output_name OUTPUT_NAME]

   Use transformation matrices generated by chim_session_to_mtx.py to populate a
   template relion star file Euler angles and rotation origins.

   optional arguments:
     -h, --help            show this help message and exit
     --template_star TEMPLATE_STAR
                    RELION *.star file to use as template.
     --chimera_mtx_output CHIMERA_MTX_OUTPUT
                    Output file from chim_session_to_mtx.py
     --rln_particle_particle_csv RLN_PARTICLE_PARTICLE_CSV
                    A "coma separated values" file denoting equivalences
                    between the naming of particles in the *.star file and
                    chimera sessions. E.g., if a particle is named "Partic
                    les/Tomograms/tomogram101/tomogram101_subtomo000004.mr
                    c" in the star file, and "tomo1_4.mrc" in the Chimera
                    session, the *.csv file should have a line with the
                    following format:
                    "tomogram101_subtomo000004.mrc,tomo1_4.mrc". There
                    should be one pair per line.
     --angpix ANGPIX       Particle pixel size.
     --output_name OUTPUT_NAME
                    Name for output *.star file

Here's a commandline example:

python ./mtx_to_star.py --template_star you_template.star --chimera_mtx_output matrix_dictionary.txt --rln_particle_particle_csv equiv.csv --angpix 10.88

Usage of script for shifting particles according to the final reconstruction

Activate the py37 environment using conda activate py37 to run shift_particles_starfile.py.

python ./shift_particles_starfile.py --help

It will generate the following output:

usage: shift_particles_starfile.py [-h] --template_star TEMPLATE_STAR
                               [--angpix ANGPIX] --translation TRANSLATION
                               --particle_box_size PARTICLE_BOX_SIZE
                               [--output_name OUTPUT_NAME]

Use transformation matrices generated by chim_session_to_mtx.py to populate a
template relion star file Euler angles and rotation origins.

optional arguments:
  -h, --help            show this help message and exit
  --template_star TEMPLATE_STAR
                    RELION *.star file containing particles to be shifted.
  --angpix ANGPIX       Particle pixel size.
  --translation TRANSLATION
                    Desired translation in pixels: X,Y,Z
  --particle_box_size PARTICLE_BOX_SIZE
                    Particle box size in pixels: X,Y,Z
  --output_name OUTPUT_NAME
                    Name for output *.star file. Otherwise it will be
                    called "<original name>.shifted.star".

Here's a commandline example:

python ./shift_particles_starfile.py --template_star skeletondata.init_from_Chimera.star --angpix 8.52 --translation 0,0,-15 --particle_box_size 96,96,96

Usage of UCSF Chimera fitmap iterator

1. First, activate the pychimera environment using conda activate pychimera to run chimera_fitmap_search.py. You can print a help message typing:

python ./chimera_fitmap_search.py --help

It will generate the following output:

usage: chimera_fitmap_search.py [-h] --subtomogram_path SUBTOMOGRAM_PATH --reference_path
               REFERENCE_PATH --s_angpix S_ANGPIX --r_angpix R_ANGPIX
               --s_level S_LEVEL --r_level R_LEVEL

This script uses the UCSF Chimera fitmap function to generate sessions with a
reference density of interest docked in a series of subtomograms. We recomend
to first open a few subtomograms along the reference to evaluate ideal
isosurface countour levels. See Volume Viewer > Level in the UCSF Chimera
graphical interface.

optional arguments:
  -h, --help            show this help message and exit
  --subtomogram_path SUBTOMOGRAM_PATH
                    Absolute path to directory where all subtomograms are
                    stored, e.g., "/User/sessions/*.mrc"
  --reference_path REFERENCE_PATH
                    Absolute path to reference density. E.g.,
                    "./low_passed_ribosome.mrc"
  --s_angpix S_ANGPIX   Pixel size of subtomograms
  --r_angpix R_ANGPIX   Pixel size of references
  --s_level S_LEVEL     Subtomogram isosurface contour level
  --r_level R_LEVEL     Reference isosurface contour level

Here's a commandline example:

python ./chimera_fitmap_search.py --subtomogram_path ./tomogram1_subtomo000007_gaussian.mrc --reference_path ./dynactin_resampled.8p52.mrc --s_angpix 8.25 --r_angpix 8.25 --s_level 1.25 --r_level 0.036

If you are interested in only running the fitmap iterator, and you already have the chimera command line interpreter, we provide a version of chimera_fitmap_search.py (chimera_fitmap_search_quick.py) that you can run using, for example:

chimera --nogui --script "chimera_fitmap_search_quick.py --subtomogram_path ./tomogram1_subtomo000007_gaussian.mrc --reference_path ./dynactin_resampled.8p52.mrc --s_angpix 8.25 --r_angpix 8.25 --s_level 1.25 --r_level 0.036"