F - Image and data-handling

For full functionality, CC needs for each image several files:

1. The file for image 1 (channel 1)
2. Possibly the files for images (channels) 2 or 3 (can be deactivated)
3. The LabelMatrixImage (=segmented image) of the current image (see below)
4. The ImData.mat file (see below)
5. The _handles.mat file

Files for images (channels) 1, 2 and 3 can be any files accepted by CellProfiler, usually .tiff files. Enhanced CellClassifier assumes all files to be in one folder.

Normal 0 false false false EN-US X-NONE X-NONE st1\:*{behavior:url(#ieooui) } /* Style Definitions */ table.MsoNormalTable {mso-style-name:"Table Normal"; mso-tstyle-rowband-size:0; mso-tstyle-colband-size:0; mso-style-noshow:yes; mso-style-priority:99; mso-style-qformat:yes; mso-style-parent:""; mso-padding-alt:0in 5.4pt 0in 5.4pt; mso-para-margin:0in; mso-para-margin-bottom:.0001pt; mso-pagination:widow-orphan; font-size:10.0pt; font-family:"Times New Roman","serif";}

The LabelMatrixImage=Segmented Image gives information about the positions of the objects to be classified. It is an image as big as the original image with the shapes of the identified objects, stored as a .mat file. It has to be generated by CellProfiler with the ConvertToImage module (see preparations). The files for all LabelMatrixImages of one plate need to be in one folder.

The ImData.mat file contains all CellProfiler measurements for one image. CellProfiler saves the results of its analysis either in one large _OUT.mat file or, if run on a Linux-cluster in several smaller _OUT.mat files for a couple of images (batch of images). While "splitting data" CC loads these files, merges them and stores them as .mat files on a per image basis as an _ImData.mat file (see below). This is necessary since with 384-well plates or with extensive measurements loading of data might take several minutes and often causes Matlab: Out of Memory errors.

The _handles.mat file contains further information, for instance about the CellProfiler analysis and image file names.

Merge and Split Datafiles

File >> Merge and Split Datafiles >> Merge single CP.Outputfile

File >> Merge and Split Datafiles >> Merge and Split Cluster-Datafile

Please browse to the folder where the CellProfiler outputfiles are located. Simply select one Matlab file in this folder. CC will now generate an _Image_Data folder in this directory. Now all CP Output files are loaded and saved on an image per image basis as ImData.mat files. After the last file is loaded, CC performs a sanity check (for completeness) and saves the _handles.mat file with general information.

File >> Merge and Split Datafiles >> Merge and Split Cluster-Datafiles in Directory

If you are planning to merge several CP-Output folders the above mentioned procedure can be tiresome. If all files are located in one master folder you can select "Datafiles in directory" and browse to the master folder. All output folders in this master folder will be split.

New Experiment (new plate)

>> File >> New Experiment (new plate)

When loading data from a new plate, all trained cells will be deleted (please save Cell data before this step). Deleting data is necessary to avoid data corruption since measurements might be non compatible. You will be asked to confirm deleting data. A new graphical user interface is opened. You need to select 4 files/ folders:

1. CellProfiler Output File: the _handles.mat file of the CellProfiler analysis (see above). Please browse to the respective file
   Note: CC facilitates the selection, it is sufficient to select any file within this folder, or even any file in the original CP-folder, for instance the _data.mat file of the cluster analysis, which should be in the folder above.
2. Path for Original Images: The same for the Original Images.
3. Path for Segmented Images: The LabelMatrixImages (see above). Please browse to the respective folder or copy the path into the respective window.
   LUCKY: CC loads the CP-output file selected and checks the folders you have been using. If you have not moved your images after the CellProfiler analysis, CC will find them (if you are lucky…)
4. LibSVM Path: In order to classify cells, CC uses libsvm. CC tries to locate this folder automatically. If this does not work, you can specify the respective folder here (or in the settings file).

Show Data and Image Directories

>> File >> Show Data and Image Directories

The 4 directories mentioned in "New Experiment" are displayed.