Context-explorer

Context-explorer is a software program that facilitates analyses of data extracted from microscope images of cells.

Motivation

The analyses methods in context-explorer focuses on how populations of cells are affected by their microenvironment, including local variations in cell signalling and cell density. It is currently difficult for scientist without previous programming experience to quantify these variables, although it is relevant for many research areas. Facilitating this type of analyses can help scientist around the world to improve their understanding of cellular behavior and accelerate their research.

Graphical overview

_images/fig1-overview.png

Installing Context-explorer

The fastest way to get up and running, is to install context-explorer via the conda package manager from the Anaconda Python distribution:

  1. Download and install Anaconda Python for your platform.

2a. Open the Anaconda navigator and search for context-explorer

2b. Open a terminal or the Anaconda Prompt and type conda install -c joelostblom context-explorer.

Using context-explorer

Launch the context-explorer by opening a terminal/the Anaconda Navigator, typing ctexplorer, and hitting enter. Keep the terminal window open, there is where context-explorer will output messages of what it is doing.

Input file

The minimum required input is a CSV-file containing single cell measurements including xy-coordinates and field number for each cell. Such measurements can be extracted from microscope images using software such as CellProfiler or vendor-specific alternatives. The field number represents the position of the tile/site within each well and the xy-coordinates are the position of each cell within these tile. Ideally, at least one additional measure of interest is present, e.g. fluorescent intensity or nuclear area.

Workflow overview

_images/gui.png

There is a brief video tutorial of how to use the software.

After loading the desired CSV-file, assign treatment groups via the plate layout tab. In the threshold tab, it is possible to create histograms and scatter plots of the desired measurements. One plot will be created per well. Basic column modifications (e.g., addition and deletion) can be performed via the ColumnModification tab.

In the colony identification tab, cells can be grouped into clusters based on density variations within each well. The clustering parameters can be adjusted interactively, so that the clusters represent biologically relevant entities, e.g. colonies of stem cells or micropatterned cells. This grouping is helpful to assess the variation of the cellular microenvironment on cell fate through comparisons between groups at distinct spatial locations within the clusters, .e.g. comparing the expression of certain proteins between cells growing at the colony edge vs the middle of the colony.

The data can then be saved via the SaveData tab either as single cells data or aggregated at the level of choice (e.g. average measurements per well or treatment group).

Package build instructions

These instructions are not needed if you just want to use context-explorer. The steps below details how to build a Python package (1-2) and upload it to https://pypi.org (3), and how to build a conda package (4-6) and upload it to https://www.anaconda.com/download (7).

  1. Create a git tag with the version number.

    git tag <major.minor.patch>
    

    If there are any changes in the working tree since this tag, these needs to be stashed or versioneer will create a version tag that is incompatible with PyPI.

  2. Build the source distribution (for conda to use in the next step) and at least one binary wheel (faster for pip install).

    python setup.py sdist bdist_wheel
    
  3. Upload the package to PyPI.org.

    twine upload dist/*<major.minor.path>*
    
  4. A conda build recipe can be created from the PyPI package.

    conda skeleton pypi context-explorer
    

    This will create the directory context-explorer with the file meta.yml. The only manual modification that needs to be made is to replace the occurrences of pyqt5 with pyqt, since this package is named differently on PyPI and in conda:

  5. Build a condo package from the recipe.

    conda build context-explorer
    
  6. The conda package is usually created in ~anaconda3/conda-bld/linux-64/<package_name> and can be converted to support all platform architectures.

    conda convert --platform all <absolute/path/to/package> -o <outputdir>
    
  7. Each platform’s package needs to be uploaded separately to anaconda.org.

    anaconda upload <path/to/package>
    

    The initial package created with conda build can be uploaded automatically by setting the following.

    conda config --set anaconda_upload yes
    

Indices and tables

Heading 1

Heading 2