mmqt

Microglia morphology quantification tool (MMQT)

Toolbox for the extraction of morphological features of microglia cells from confocal microscopy image stacks containing two channels:

• anti-Iba1 staining of microglia
• DAPI staining of nuclei

The software will segment foreground from background, create a skeleton of the microglia, segregate individual microglia cells and extract morphological features.

For details see the method paper: Heindl S, Gesierich B, Benakis C, Llovera-Garcia G, Duering M, Liesz A
Automated Morphological Analysis of Microglia After Stroke
Front. Cell. Neurosci. 2018
https://www.frontiersin.org/articles/10.3389/fncel.2018.00106

This toolbox was developed at the Institute for Stroke and Dementia Research (ISD), Munich, Germany. Please read the LICENSE file before using MMQT.

Content

• Hardware and software requirements
• Getting started
• Accepted data formats
• Output files

Hardware and software requirements

• Modern CPU, e.g. Intel Core i7
• at least 16 GB RAM
• Obligatory:
  • MATLAB® R2016B (The Mathworks, newer versions untested but might work) with the toolboxes:
  • Image Processing Toolbox 9.5
  • Statistics and Machine Learning Toolbox 11.0
• When using Carl Zeiss Image Data File (CZI-file; file-extension .czi) as input::
  Bio-Formats software tool for MATLAB available from the Open Microscopy Environment consortium.

Getting started

• Add the MMQT folder to the Matlab path
• Try one of the example scripts in the subfolder ./example
• Example CZI data can be downloaded from the ISD Research Server

Accepted data formats

The MMQT toolbox needs confocal microscopy Z-stack image data with two channels:

• Channel 1: DAPI staining of nuclei
• Channel 2: anti-Iba1 staining of microglia

Images can be provided in two file formats:

• Carl Zeiss Image Data File (CZI-file; file-extension .czi).
  For details on how to import them see example script mmqt_example_CZI.m.

• MATLAB® formatted binary file (MAT-file; file-extension: .mat).
  MAT-files have to contain two variables:

  • img: 4D array containing the raw image data with 8 bit integer encoding; dimensions should correspond to X/Y/Z directions and the color channels.
  • hdr: structure with two fields, describing dimensionality and resolution (given as scaling factors) of the raw data in the array "img":
    • hdr.dim: vector of length 4, with values specifying the number of dimensions in array "img" and the size of each dimension.
      For example, hdr.dim=[4,1024,1024,132,2] specifies that the array "img" has 4 dimensions with the size of dimensions being respectively 1024, 1024, 132 and 2.
    • hdr.pixdim: vector of length 4, with values specifying the number of dimensions in array "img" and the factors to scale each dimension to micrometer. The forth dimension corresponds to the color channels and therefore has no scaling applied to it. Therefore, set its scaling factor to 1.
      For example, hdr.pixdim=[4,0.2076,0.2076,0.4,1] specifies that the array "img" has 4 dimensions and the X/Y dimension has to be scaled by 0.2076 and the Z dimension by 0.4.

Please note that scaling information for the spatial dimensions is crucial for correct processing of the image stacks. This scaling factors sould result in a micrometer scale. The MMQT software will try to automatically extract this information from CZI-files. If you work with MAT-files, make sure that this information is set correctly in this file (see above).

Output files

The output files creates by MMQT are documented in the following files: mmqt_output_files.xlsx (MS Excel file) and mmqt_output_files.txt (tab-separated text)

The scores for the extracted shape-features are stored in a tab-separated text-file with the suffix _features_summarized.txt
The ID of a cell, as given in this text-file, corresponds to the value of the voxels, which belong to the same cell, in the volumetric dataset being saved in the file with suffix _stack5_segregated_cells.mat.