tttsa

Automated tilt-series alignment for cryo-ET using teamtomo-torch!

The implementation follows the algorithm in the AreTomo paper!

Installation

pip install tttsa

Example

The program is currently not optimal in memory usage, so it's advised to run with an 8x downsampled (binned) tilt-series.

import torch
from cryotypes.projectionmodel import ProjectionModel
from cryotypes.projectionmodel import ProjectionModelDataLabels as PMDL
from tttsa import tilt_series_alignment
from tttsa.back_projection import filtered_back_projection_3d

tilt_angles = list(range(-51, 54, 3))
n_tilts = len(tilt_angles)
# this angle is assumed to be a clockwise forward rotation after projecting the sample
tilt_axis_angle_prior = -88.7
alignment_pixel_size = 10  # in A
# calculate sample thickness in pixels, 1600 A (160nm) thick ice layer
# like in AreTomo this parameter is very critical for alignment quality
align_z = int(1600 / alignment_pixel_size)
recon_z = int(2400 / alignment_pixel_size)
weighting = "hamming"  # weighting scheme for filtered back projection
device = "cuda" if torch.cuda.is_available() else "cpu"

# create a mock tilt-series
size = 400
tilt_series = torch.rand((n_tilts, size, size))

# Initialize the projection-model prior
projection_model_prior = ProjectionModel(
    {
        PMDL.ROTATION_Z: tilt_axis_angle_prior,
        PMDL.ROTATION_Y: tilt_angles,
        PMDL.ROTATION_X: 0.0,
        PMDL.SHIFT_X: 0.0,
        PMDL.SHIFT_Y: 0.0,
        PMDL.EXPERIMENT_ID: 'position_1',
        PMDL.PIXEL_SPACING: alignment_pixel_size,
        PMDL.SOURCE: 'position_1.st',
    }
)

projection_model_optimized = tilt_series_alignment(
    tilt_series.to(device),
    projection_model_prior,
    align_z,
    find_tilt_angle_offset=True,
)

final = filtered_back_projection_3d(
    tilt_series,
    (recon_z, size, size),
    projection_model_optimized,
    weighting=weighting,
)

Give it a whirl?

This will run the example (after downloading and installing the repository):

cd examples
python usage.py

It will take some time to initially download the data from Zenodo. The reconstruction is stored in examples /data/tomo2000528_107.mrc.