Subtomograms

This repository implements the Binary Decomposition and Golden Ratio strategies for performing computed tomography of dynamic processes described in Kaestner A. et al. Optical Engineering 50(12), 123201 (December 2011) (https://doi.org/10.1117/1.3660298).

Installation

Note: you may need to use python3 instead of python, depending on your system.

Currently only working as a local import. Make sure you run python from the same directory as this repo.

$ git clone git@github.com:verejoel/subtomograms.git
$ cd subtomograms

Create a Virtual Environment

Do this according to your favourite manager (direnv, pyenv, pipenv). For plain python:

$ python -m venv ./venv
$ source venv/bin/activate

Install dependencies

$ pip install -r requirements.txt

Usage

Start a python interpreter session:

$ python

In the interactive session, first import the Generator classes:

>>> from generator import GoldenGenerator, BinaryGenerator

You can now use the GoldenGenerator and BinaryGenerator classes to build subtomograms.

Binary Decomposition

To use binary decomposition, initialise an object in the following way:

>>> bg = BinaryGenerator(4, 10)
Initialised!
Total projections: 40
Step size per subtomogram: 18.0
Total step size: 4.5

Here, the numbers 4 and 10 correspond to the number of subtomograms, and the number of projections per subtomogram, respectively. The object will be initialised, and report the total number of projects, step size per subtomogram, and the total step size when all subtomograms are combined.

You can now calculate the angles for each subtomogram:

>>> bg.print_angles()
Subtomogram     Angles
0               [  0.  36.  72. 108. 144. 180. 216. 252. 288. 324.]
1               [ 18.  54.  90. 126. 162. 198. 234. 270. 306. 342.]
2               [  9.  45.  81. 117. 153. 189. 225. 261. 297. 333.]
3               [ 27.  63.  99. 135. 171. 207. 243. 279. 315. 351.]

By default, the projections are distrubuted over a full circle. To distribute across a half-circle, call the generator with the full_circle flag set to False:

>>> bg = BinaryGenerator(4, 10, full_circle=False)
>>> bg.print_angles()
Subtomogram     Angles
0               [  0.  18.  36.  54.  72.  90. 108. 126. 144. 162.]
1               [  9.  27.  45.  63.  81.  99. 117. 135. 153. 171.]
2               [  4.5  22.5  40.5  58.5  76.5  94.5 112.5 130.5 148.5 166.5]
3               [ 13.5  31.5  49.5  67.5  85.5 103.5 121.5 139.5 157.5 175.5]

Changing the output units

By default, the generators output angles in units of degrees, as this is typically more useful for experiments. Should you prefer radians, simply initialise the generator with the degrees flag set to false:

>>> bg = BinaryGenerator(4, 10, degrees=False)
>>> bg.print_angles()
Subtomogram     Angles
0               [0.    0.628 1.257 1.885 2.513 3.142 3.77  4.398 5.027 5.655]
1               [0.314 0.942 1.571 2.199 2.827 3.456 4.084 4.712 5.341 5.969]
2               [0.157 0.785 1.414 2.042 2.67  3.299 3.927 4.555 5.184 5.812]
3               [0.471 1.1   1.728 2.356 2.985 3.613 4.241 4.869 5.498 6.126]

By default, values are rounded to three decimal places - to override this, set the class variable DP after you create the generator:

>>> bg.DP = 5
>>> bg.print_angles()
Subtomogram     Angles
0               [0.      0.62832 1.25664 1.88496 2.51327 3.14159 3.76991 4.39823 5.02655 5.65487]
1               [0.31416 0.94248 1.5708  2.19911 2.82743 3.45575 4.08407 4.71239 5.34071 5.96903]
2               [0.15708 0.7854  1.41372 2.04204 2.67035 3.29867 3.92699 4.55531 5.18363 5.81195]
3               [0.47124 1.09956 1.72788 2.35619 2.98451 3.61283 4.24115 4.86947 5.49779 6.12611]

verejoel / subtomograms