• This tutorial features images of human induced pluripotent stem cells from the Allen Institute of Cell Science. More details are available at the following link: https://bbbc.broadinstitute.org/BBBC034.
• CellProfiler 3D currently only works with TIFF files. TIFF files can be rather complicated, having hyper-stack structures with all channels and z-planes in a single file. The acceptable CellProfiler format for storing z-stacks is to have a separate TIFF file for each channel.
• CellProfiler can be used to convert from other file formats to individual TIFF files for each channel using the SaveImages module.
• Note that this tutorial is an advanced tutorial. We recommend completing the Translocation tutorial in order to learn principles of image thresholding and segmentation prior to starting this tutorial.
• Helpful video tutorials are available on the Center for Open Bioimage Analysis YouTube page at https://www.youtube.com/channel/UC_id9sE-vu_i30Bd-skay7Q.

1. Highlight the Images module.
2. Drag-and-drop the images you will analyze into the Images module window.
3. Highlight the Metadata module.
4. Enter the following regular expression ^(?P<Plate>.*)_xy(?P<Site>[0-9])_ch(?P<ChannelNumber>[0-9]). This regular expression will parse the filenames and organize the data.
5. Highlight the NamesAndTypes module.
6. Assign a name to "Images matching rules".
7. Choose “Process as 3D”
8. Populate the fields for "Relative Pixel Spacing".
   • Fiji > Image > Show Info… (Ctrl + I)
   • Search for something like “Voxel size” or record this metadata when collecting your own images
   • The actual units do not matter, rather their relative proportion. The numbers are unitless and therefore the decimal place does not matter.
   • For this example, the relative pixel spacing is 0.065 in x and y and 0.29 pixels in z.
9. Create "rule criteria" to identify an image by its color/channel. For example, using the Metadata you just extracted - Metadata -> Does -> Have ChannelNumber matching -> 0 would match the first image.
10. Give the images "variable names" that describe the contents in the image. For example, use the name dna or dapi to describe an image stained with DAPI.
11. Add images with rulesets for the other channels in the experiment. In this case, Channel 0 contains images of the plasma membrane, Channel 1 contains images of mitochondria, and Channel 2 contain images of DNA.

Before attempting to segment the cells in the images, conditioning the images with filters and various image processing methods will improve the results.

1. Add a RescaleIntensity module for the DNA channel. Rescaling the DNA image proportionally stretches the intensity values to the full intensity range, from 0 to 1. In this case, we find that rescaling improves the thresholding and subsequent segmentation of nuclei. When using rescaling in your pipelines, be careful to perform measurements on the original images, not the rescaled images.

   

2. Add a Resize module. Processing 3D images requires much more computation time than 2D images. Often, downsampling an image can yield large performance gains and at the same time smooth an image to remove noise. Final segmentation results will be minimally affected by downsampling if the objects of interest are relatively large compared to the pixel size. Choose a value of 0.5.

   

3. Add a MedianFilter module. A median filter will homogenize the signal within the nucleus and reduce noise in the background. DNA is not uniformly distributed throughout the nucleus, which can lead to holes forming in the downstream object identification. A median filter will preserve boundaries better than other smoothing filters such as the Gaussian filter. Choose a filter size of 5. This number was chosen empirically: it is smaller than the diameter of a typical nucleus; it is small enough that nuclei aren't merged together, yet large enough to suppress over-segmentation of the nuclei.

   

1. Add an Threshold module. This identifies a pixel intensity value to separate the foreground (nuclei) from the background. Empirically, we've found that a two-class Otsu threshold works well for this data. We encourage you to try other thresholding methods to compare the outputs.

   

2. Add a RemoveHoles module. This module implements an algorithm that will remove small holes within the nucleus. Any remaining holes will contribute to over-segmentation of the nuclei. Choose a size of 20.

   

3. Add a Watershed module. This module implements the watershed algorithm, which will segment the nuclei. Select a Footprint of 10 and Downsample by 2. Downsampling reduces processing time and decreases noise. For more information on the watershed algorithm refer to this helpful MATLAB blog post.

   

4. Add a ResizeObjects module to return the segmented nuclei to the size of the original image. Since the original image was scaled down by 0.5, it must be scaled up by 2. The output of this module is the nuclei we are seeking, so name these objects accordingly, e.g. Nuclei.

   

Now that we've segmented the nuclei we want to segment the cytoplasm for each nuclei whose boundaries are defined by the membrane channel. The membrane channel presents more of a challenge, because unlike the nuclei, the membrane signal is variable and the boundaries are connected together in a sort of mesh. This challenge is mitigated by the fact that the location of the nuclei can be used to help identify regions with cells.

1. Shrink the nuclei to make them more seed-like by adding an ErodeObjects module. Use the ball structuring element with a size of 5. Select "Yes" for the "Prevent object removal" option in order to avoid losing any nuclei.

   We've found that we can achieve the best results by applying ErodeObjects to the output of the Watershed module rather than the resized Nuclei that are at the original size (since the Watershed output has been downsampled, the resulting seeds from ErodeObjects are smaller and more seed-like).

   

2. Resize these eroded objects using the ResizeObjects module with a factor of 2.

   

3. Next convert the eroded and resized nuclei to an image using the ConvertObjectsToImage module. Select the uint16 color format. This image will serve as the seeds for segmenting the cells.

   

The Watershed module finds objects that have bright signal, so the cytoplasm that will define the cell volume should have bright signal. However, this is not the case in the membrane channel; it must be transformed into an image where the cytoplasm is bright and the boundaries between the cells are dark. Therefore, we will invert the membrane channel to achieve this effect.

1. Add a Threshold module and threshold the rescaled membrane image. We find that the Otsu three-class method with middle intensity pixels assigned to the foreground works well, but feel free to try others.

   

2. Add an ImageMath module. Within the ImageMath module choose the Invert operation, and invert the thresholded membrane.

   

   We invert the thresholded membrane in order to create a binary image where the pixels inside of cells are bright (1) and the pixels surrounding cells are black (0).

3. Add a RemoveHoles module to remove the small holes in the segmentation of the cell interior. This helps to prevent the cells from being split during the Watershed segmentation. Choose a size of 20. This result will be referred to as the Inverted Membrane.

   

   We cannot use the inverted membrane image as the cytoplasm just yet. The space above and the below the monolayer is also of high signal. The Watershed module cannot distinguish that this is not cytoplasm, so it will have to be removed. To do this we will take advantage of the signal across all channels to define the boundaries of the monolayer.

4. Add another ImageMath module. Add all of the original images together. This creates a composite image that will be used to define where cells are present and the background above and below the cells. This image will be referred to as the Monolayer

   

5. Add a Resize module to resize the Monolayer with a Resizing factor of 0.25. Downsampling the image makes processing faster and decreases noise.

   

6. Add a Closing module. Choose a size of 17 to blend the signal together. The result should look like a cloud of signal where the monolayer resides.

   

7. Add a Resize module to resize the closed Monolayer back to its original size, using a Resizing factor of 4.

   

8. Add a Threshold module and threshold the smoothed monolayer image. This will define what is and is not monolayer. Note that the space above and below the monolayer is primarily black.

   

   Now we will combine the information from the membrane channel with what we identified as the monolayer. We will do this by using the MaskImage module to apply the MonolayerMask to the thresholded membrane.

9. Add a MaskImage module. You will use an Image as a mask (the MonolayerMask image generated in the previous step). In this case, the mask does not need to be inverted. Note that the planes on the bottom and top of the z-stack are black in the masked image.

   

10. Add an ErodeImage module. We will use this module to erode the membrane image generated in the previous step. Eroding using a ball of size 1 improves the separation between individual cells in the Watershed segmentation (the next step).

    

11. Add a Watershed module. The input is the result of the previous ErodeImage module, referred to here as the MembFinal. Change the Generate from option to Markers. The Markers will be the NucleiSeeds image, which is the output of the ConvertObjectsToImage module. Finally, set the Mask to also be the MembFinal. This will help preserve the cell boundaries.

    

Now that the nuclei and cells have been segmented in this monolayer, measurements can be made using modules from the Measurements category.

1. Add any desired measurements modules. For example, you might choose to MeasureObjectIntensity and/or MeasureObjectSizeShape. When applying these measurements, be careful to measure the original images, not rescaled images.

   

Congratulations! The nuclei and cells have been segmented and measured in this monolayer. Visuals that reveal the details of the segmentation can be also be created within CellProfiler. The following steps will walk through two different options to visualize your CellProfiler segmentations.

1. The OverlayObjects module will overlay the objects as colored masks on the image. We recommend overlaying onto rescaled images, which will be easier to visualize outside of CellProfiler. For example, you can choose the Nuclei as the objects and the RescaledDNA as your image. These are useful for visualization, but unfortunately cannot be saved.

   

2. You can also convert the objects to images using the ConvertObjectsToImage module and then save the output using SaveImages. This option will allow you to visualize the segmentations directly in Fiji.

   

   After running these last two modules an output image will be created and saved to the output directory. Use Fiji to inspect the this image.

1. Save the output of the measurements modules using ExportToSpreadsheet or ExportToDatabase.

It's good practice to place all export modules at the end of your pipeline. CellProfiler automatically calculates execution times for each module that was run before the export module. By placing your export modules at the end of your pipeline, you will have access to module execution times for each module in your pipeline.

Thank you for completing the 3d monolayer tutorial!