acryo

acryo

acryo is an extensible cryo-EM/ET toolkit for Python.

The purpose of this library is to make data analysis of cryo-EM/ET safer, efficient, reproducible and customizable for everyone. Scientists can avoid the error-prone CLI-based data handling, such as writing out the results to the files every time and manage all the result just by the file names.

📘 Documentation

Install

Use pip

pip install acryo -U

From source

git clone git+https://github.com/hanjinliu/acryo.git
cd acryo
pip install -e .

Features

1. Out-of-core and parallel processing during subtomogram averaging/alignment to make full use of CPU.
2. Extensible and ready-to-use alignment models.
3. Manage subtomogram loading tasks from single or multiple tomograms in the same API.
4. Tomogram and tilt series simulation.
5. Masked PCA clustering.

Code Snippet

import polars as pl
from acryo import SubtomogramLoader, Molecules  # acryo objects
from acryo.tilt import single_axis  # missing wedge model
from acryo.pipe import soft_otsu  # data input pipelines

# construct a loader
loader = SubtomogramLoader.imread(
    "path/to/tomogram.mrc",
    molecules=Molecules.from_csv("path/to/molecules.csv"),
)

# filter out bad alignment in polars way
loader_filt = loader.filter(pl.col("score") > 0.7)

# averaging
avg = loader_filt.average(output_shape=(48, 48, 48))

# alignment
aligned_loader = loader.align(
    template=avg,                           # use the average as template
    mask=soft_otsu(sigma=2, radius=2),      # apply soft-Otsu to template to make the mask
    tilt=single_axis((-45, 45), axis="y"),  # range of tilt series degrees.
    cutoff=0.5,                             # lowpass filtering cutoff
    max_shifts=(4, 4, 4),                   # search space limits
)