Tglow: Core Python component of the tglow imaging pipeline

tglow-core is the Python core component of the Tglow high-content imaging (HCI) analysis pipeline. It provides utilities to index and read multi-well plate images and parsers for PerkinElmer (Opera Phenix / Operetta) exports. The package is used by the tglow-pipeline workflows to load, preprocess and write OME-TIFF images arranged in the common /plate/row/col/field.ome.tiff (CYZX) layout.

Key features

• Read and write CYZX / ZYX / YX image arrays via AICSImageReader / AICSImageWriter (wrappers around aicsimageio)
• Parse Revity/PerkinElmer Index.xml exports (PerkinElmerParser) and convert to a simple, Python-friendly index
• Convert large Revity/PerkinElmer exports to a much lower number of /plate/row/col/field.ome.tiff files
• Index and query plate/row/col/field image layouts using an ImageQuery object
• Utilities for registration, flatfield correction and numeric conversions designed to work with tglow-pipeline

Installation

I recommend installing the published PyPI release where possible:

pip install tglow-core

To install the latest development version from the repository (editable install):

git clone https://github.com/TrynkaLab/tglow-core
cd tglow-core
pip install -e .

Basic usage

Build an index from a PerkinElmer export and read a single image:

from tglow.io.tglow_io import PerkinElmerRawReader
from tglow.io.image_query import ImageQuery

reader = PerkinElmerRawReader('path/to/Index.xml', '/data/exports')
iq = ImageQuery.from_plate_well('plate1', 'A01')
image = reader.read_image(iq)  # returns a numpy array

Read and write an OME-TIFF stack organized by plate/row/col/field:

from tglow.io.tglow_io import AICSImageReader, AICSImageWriter
from tglow.io.image_query import ImageQuery

reader = AICSImageReader('/data/plates')
writer = AICSImageWriter('/output/plates')
iq = ImageQuery('plate1', 1, 1, 'field001')
stack = reader.read_stack(iq)
writer.write_stack(stack, iq)

Notes and migration to BioIO

• This package currently wraps aicsimageio. As that project has been superseded by newer tooling, consider migrating to bioio or equivalent in future releases.

Known issues

There is a known issue with BaSiCpy (https://github.com/peng-lab/BaSiCPy/issues/162). This requires using specific, older versions of hyperactive and gradient-free-optimizers, which can in turn require an older pandas version. The same goes for aicsimageio. The dependency chain can be inconvenient; I'll update this project and migrate to BioIO libraries as newer releases become available.

Acknowledgements

• Martin Prete: initial XML parsing code adapted for this project

References

• https://github.com/AllenCellModeling/aicsimageio
• https://github.com/bioio-devs/bioio