Step-by-step example

Here, we propose step-by-step examples of possible usage of the pipeline. You can download test data from this zenodo repository: https://zenodo.org/records/17048217 to follow it.

All the examples here are done with the image AB3-HG-AQUCISITION-4CHANNELS-SHRNActrl-filtered_minicrop.ims.

A- Count the number of smFish spots by cell

A1 - Open the image and check/set the metadata

Open Napari and start FishFeats with Plugins>FISHFEATS>Start fishfeats.

A window dialog appears to select the image to analyse. Browse your folders and select the downloaded image AB3-HG-AQUCISITION-4CHANNELS-SHRNActrl-filtered_minicrop.ims (or another one on which you want to do this test).

The image is loaded and is displayed with its color channels side by side. At the right of the interface, you can set the metadata of the image.

Check that the scaling metadata were correctly loaded and correct them if needed (see image below).

Then, check that the 3D direcion of the image is correct (the higher Z slices correspond to the apical cells = top high z).

Set the channels number that contains the junction staining (Zo1) and/or the nuclei staining (DAPI, PCNA, Sox2). In our example image, the first channel contains the junction (Zo1), which correspond to number 0 (originalChannel0), in red. The nuclei channel is the last channel, originalChannel3.

Finally click on Update to set these properties for the image. All this set-up will be saved in the configuration file AB3-HG-AQUCISITION-4CHANNELS-SHRNActrl-filtered_minicrop.cfg associated with this image. When you reload this image in FishFeats, it will read this configuration file and load it back. This file can also be useful if you don't remember which option and parameters you selected for a given step. You can open this file and you will see the list of steps and parameter values used.

You can then choose which steps to perform depending of your analysis by selecting the step in the Main panel in the right side of the window.

A2 - Segment epithelia cells

To segment the apical contour of the cells, choose Get cells option.

If you have already done this step, the plugin will write Found projection file and/or Found cell file if it found the corresponding files with the default names. You can then choose the option Load previous files to directly load them and go to the manual curation step.

Else, select do projection and segmentation.

An interface opens to perform the projection. You can either load a file of the projected junction staining if ou have done it with another external software (e.g. CARE, LocalZProjector), by clicking on choose file.

Otherwise, click on Project now to calculates the local projection with FishFeats. The projected results will be overlaid in white in your window (image below). If you are not satified with the results, you can click on Advanced to change the parameters and Project now again.

Then click on Projection done to perform the segmentation on the projected channel.

Choose Epyseg in the Method interface to perform cell apical segmentation (or CellPose). If you haven't installed Epyseg (not mandatory), you will get an error message. Install it in your virtual environement by typing pip install epyseg.

Click on segment cells and wait for the segmentation to be calculated (0.105 min on this test image with 1 GPU).

You can manually correct the segmentation with shortcuts that are indicated at the top left of your window.

Click on Cells done to finish this step and save the results. Wait while the program save the data and replace the projected cells in 3D. This can take a little time as it is looking for the best Z-position for each cell.

You have now segmented all the apical cell contours and estimated their position in 3D. If you open the result file (AB3-HG-AQUCISITION-4CHANNELS-SHRNActrl-filtered_minicrop_results.csv) in a table editor (e.g. Excel, R, Prism..) you can already analyse these data (number of cell, apical area, position). This file will be enriched with all new information that you will add by performing other steps.

A3 - Segment the smFish staining

Select Get RNA in the Main interface for this step.

The interface that opens allow you to choose the channel to segment, and the parameter for the segmentation. Choose 2 in RNA channel parameter.

You can now segment this channel with big-fish (if it is not installed (not mandatory), install it with pip install big-fish). Choose the approximate size of one dots in XY, and in Z (not the same as because of imaging technique a circular spot is usually a rugby-balloon shape). Put 1200 nm for Spot Z radius and 400 nm for Spot X,Y radius. Select Automatic for the threshold parameter, to let big-fish select the intensity threshold automatically. If the automatic threshold does not good result, you can then unclick it and the value determined by big-fish will be written in the parameter box. You can then modify it.

Click on Get RNA channel to perform the segmentation with the selected parameters.

Wait for the segmentation to be calculated (0.103 min on the test image).

When it is finished, a new layer is added to the view, called assignedRNA2. The center of each RNA spot is displayed as a Napari point in white. Two new panels have been added to the right side interface of the RNAs option. You can change the points size with the interface in the right side, as well as colored them by specific properties, in the Point display part in EditRNA2 panel.

A4 - Assign the segmented dots to their cell, count them

Automatic assignement options allow to perform an automatic assignement of each spot to its most likely cell. We will do assignement with the simplest method, which is the Projection. Each spot will be assigned to the apical cell that it belongs to after an orthoogonal Z-projection.

Select Projection in Assignement method and click on Apply assignement.

When the calculation is finished (0.004 min on our test image), the points are displayed with their assigned cell color. The contour of the cell are also overlaid to see directly the correspondance.

You can correct this assignments with the available shortcuts (see text in the top-left part of the window) and with the Point correction panel.

You can save the results while you are performing the correction by clicking on Save RNAs and click on Save and quit RNA 2 when all is done.

In the results table, you will now have a new column with the counts of RNA 2 in each cell. You can also display directly these counts by clicking on Draw RNA 2 counts in the EditRNA2 panel.

You obtain the map of the cells, where each cell is colored by its number of RNA2 spots assigned to it by Projection, then manually corrected.

B - Classify the cells by their intensity in one channel

B1 - Open the image and set metadata

Open Napari and start FishFeats with Plugins>FISHFEATS>Start fishfeats.

A window dialog appears to select the image to analyse. Browse your folders and select the downloaded image AB3-HG-AQUCISITION-4CHANNELS-SHRNActrl-filtered_minicrop.ims (or another one on which you want to do this test).

The image is loaded and is displayed with its color channels side by side. At the right of the interface, you can set the metadata of the image.

If you have done the previous example, the metadata are reloaded and already set. Click on Update and directly go the next step.

Else, follow the step A1 of the previous example.

B2 - Segment the apical cell contour

If you have done the previous example A, the cell are already segmented and should be loaded if you let load previous parameter selected. Then, you can directly go to the classify cells step.

Else, see the step A2 to perform apical cell contour segmentation.

B3 - Classify the cells by intensity

Select Classify cells in the Main interface to do this step.

It will load the cell contours so that you can classify them. In the interface, select the Add/Edit a feature panel and write the name of the feature you want to classify, e.g. Positive.

Click on Do feature to open an interface that allows you to initialize the cell classification.

Select channel projection+threshold to set an initial classification based on the projected intensity of one channel. Select 1 in channel to project parameter to perform this thresholding on the originalChannel1 channel (the green one). Set Threshold to 1.2 and click Create new feature to launch the classification.

A new layer called Feat_nameofyourfeatureCells is added to the viewer. It contains the cells colored by their class. You can easily edit this classification with the shortcuts that are listed in the top left part of the window and the options in the Edit Feat_Positive panel.

Set the current class value to 2 by pressing i. Press Ctrl+right-click on a cell that you consider as Positive to set it to the 2nd class (Positive class, in light blue).

Press i or d to set the current class value to 1 (negative cells, in maroon), and Control+right click on the cells that you want to set as Negative.

You can see the table of each cell and its value for the feature (or other features already done) in the Features table menu. Click on Update table to update it to the latest value of the new feature.

Click on Stop and Save or on Feat_Positive done when you have finished to save the classification (in the result file with the other analysis).