Morphoscript analyzes the organization level of sarcomeres in cardiomyocytes and measures their local orientation and spacings. It is developed as a Matlab application. A python jupyter notebook is also available for explaining the main concepts used by the software. It should work for the version of Matlab 2017a or higher. More details can be found in our article: https://doi.org/10.1093/bioinformatics/btab400
After downloading the project, to start the application please run the file script_example.m.
If you run the file script_example.m a window will open and you will be asked for the files you want to analyze. You can make your first try by selecting the examples provided Here are the different menus now available:
Choose the paths where you want your results to be saved (TYPES OF RESULTS?)
If the pixel size can not be read from the image metadata, a window will open asking for the pixelsize (this is the case for the test1.tif, where the pixel size to be entered is 0.06 µm).
The image is now presented to the user, if it is a colored image, one can choose the best channel for doing the analysis (if "all" is selected, the image used for the analysis will be the average of the three channels)
Three new possibilities for making the mask
- manual: the contour of cells is drawn by hand, you can add as many cells as you wish
- automatic : an automatic detection of the contour can be attempted using an entropy filter an opening to remove small non objects size (whose size can be adjusted) and a disk closing of the cell to remove holes in the mask, here also the size can be adjusted
- combination : a rough contour can be drawn on which the automatic algorithm is then run.
Check if all the files added in the section of the selected files do possess a corresponding mask necessary for a subsequent analysis.
This menu is used for measuring geometric parameters of the cell characterizing cell spreading as well as the intensity of the sarcomeric structure, 4 parameters can be visualized:
- Otsu ie the mask with an automatic threshold delineating the sarcomeres
- compactness showing the boundaries of the mask
- Bounding box of the cell mask
- Ellipse, which will show the ellipse with the same second moment as the cell's mask
The original button enables to go back to the bare display of the cell of interest. The save button will save the results. Data are saved to disk in a file named '_maski_ShapeAnalysis.mat', where '' represents the name of your image, i will be the cell number on your image. In this mat file, you will get the following variables: {'MaskPixels','OtsuPixels','MaskMicrons','OtsuMicrons','Coverage','MeanImageIntensity','MaxImageIntensity','MeanProteinIntensity','Compactness,'Roundness','Rectangularity','Elongation','Ellipsoidal','Eccentricity', 'MeanBackCellintensity'}
First, check that the pixel size is correct, if not you can easily set the appropriate one. Then 3 parameters can be adjusted for the analysis
-
Window Size : size of the interrogation window for the Fourier Transform Calculation
-
Overlap of the window
-
Threshold : which will determine if the intensity of the peak found in Fourier corresponding to the sarcomere structure is intense enough to acknowledge the presence of a real structure.
Two possibilities are available :
- Run by hand, for each window under the cell mask the intermediate Fourier images and corresponding intensity curve will be displayed which enables the user to crosscheck absolutely all the results
- Run model, will run the complete analysis automatically with no display, but the user will still have the possibility to correct some window by clicking on the edit button and selecting the window of interest, it will open the intermediate Fourier display.
After a complete analysis, the order parameter, the standard deviation on orientation, and the spacing of the sarcomeres are displayed. You can save the results and go to the next cell.
Tess Homan (t.a.m.homan@tue.nl) Hélène Delanoë-Ayari (helene.ayari@univ-lyon1.fr) Adrien Moreau ( adrien.moreau@outlook.com)
GNU General Public License v3.0