Skip to content

Project 2A: Semi-automated processing using Fiji, Ilastik, and Python

Here we will use a semi-automated pipeline to segment time-lapse data and quantify cell properties over time. We will not track cells.

We will use a set of tools that are very frequently used in image analysis:

  • Fiji, aka: ImageJ on steroids, which is a generalist user-friendly, GUI based, image analysis program offering many tools and plugins for manual and semi-automated analysis.
  • Ilastik, which is a user-friendly, GUI based tool, offering several supervised machine learning pipelines.
  • python, a generalist programming language which has become the language of choice for the image analysis community

General note: this guide has been written assuming you use a Mac or Linux Command Line. For windows Powershell see here, for command prompt see here.

Note on data

The dataset we will work with here consists of a time-lapse data of a synthetic microbial cross-feeding community consisting of two auxotrophic strains of E. coli that can only grow by exchanging Amino-Acids with each other. The two cell-types are fluorescently labeled with two different colors and grow together in a mono-layer within a 2D microfluidic growth chamber (a so-called family-machine).

Create project folders

We first create a folder on your private computer for this course. Here we will assume it is ~/I2ICourse/ but you can use any folder (just change path variables accordingly).

  • Open the command line and navigate to your home folder, then create a new folder called I2ICourse for the course:
cd ~
mkdir I2ICourse
  • Next create a folder for Project-2A ~/I2ICourse/Project2A/:
cd ~/I2ICourse/
mkdir Project2A
  • And also make a folder for the processed data (output from all steps below):
cd Project2A
mkdir ProcessedData

  • This should have created the following folder: ~/I2ICourse/Project2A/ProcessedData/:

  • Note: two useful terminal commands are:

  • ls: show current folder content (on Windows use dir)
  • pwd: show path of current folder

Download Data

Here we will download the data via the command line.

  • Navigate to the data folder:
cd ~/I2ICourse/Project2A/
  • Download the data set and unzip it using:
curl -o RawData.zip https://drive.switch.ch/index.php/s/VsWWiuaIctITWQl/download
unzip RawData.zip
  • You should now have a folder ~/I2ICourse/Project2A/RawData containing 3 tif files (hint: use ls to check!)
  • We can delete the zip-file using: rm RawData.zip

Alternative Download via browser

If you have trouble downloading via command line, you can also use your browser:

  • Open this link
  • Click Download
  • Unzip the compressed file inside your data folder.
  • The data should now be located in ~/I2ICourse/Project2A/RawData
  • Please rename the folder if needed

Data preprocessing with Fiji

We will first prepare the data in Fiji

Merging & splitting color channels

Before pre-processing we need to merge the color channels.

  • Open the individual color images (pos0-[c].tif, where c={r,g,p})
  • Image -> Color -> Merge Channels
  • Make sure Create composite is selected
  • Under C1 (red) select the [r] image
  • Under C2 (green) select the [g] image

  • Under C4 (grey) select the [p] image

  • Save on disk as pos0-merged.tif

  • use your processed data folder, e.g.: ~/I2ICourse/Project2A/ProcessedData/

Crop Image

  • Make a rectangular selection around the area of interest
  • Make sure that you make the area big enough to accommodate any remaining jitter.
  • Go to Image->Crop
  • Save image as pos0-preproc-merged.tif
  • use your processed data folder, e.g.: ~/I2ICourse/Project2A/ProcessedData/

Export data for segmentation

First we split the image into separate color channels

  • Image -> Color -> Split Channels
  • Save the separate image channels under the name pos0-preproc-[c].tif

In addition we need a combined image with the red and green channels.

  • Use Merge Channels to combine the red and green channels (do not include phase!).
  • Save on disk as pos0-preproc-rg.tif
  • use your processed data folder, e.g.: ~/I2ICourse/Project2A/ProcessedData/

Aside: Other preprocessing steps

During preprocessing you would often also do some other steps, for example:

  • Registration: i.e. aligning images between frames to compensate for stage and sample drift
  • Deconvolution: i.e. correcting for diffraction artifacts to make segmentation easier.

Unfortunately we do not have time to go into this now, but please ask us during the breaks for more information!

Segment with Ilastik

We will give a brief live-demo of how to use Ilastik, please let us know when you are at this step, so that we can get the entire group together.

You can find detailed instructions (and a movie) here.

Most important steps (see also the pdf in the Project2A repository folder):

  • Open Ilastik
  • Select Pixel Classification workflow
  • Save it in processed data folder as ‘proj2A-ilastik’
  • Go to Input Data
  • Load the data of the red-green channel.
  • Important: when adding the input data, you might have to change the axis order: double click on the axis order (e.g. zcyx) and change to tcyx.
  • Go to Feature Selection
  • Select all features
  • Go to Training
  • Make two labels: Cells and BG
  • Add sparse training points to indicate which pixels belong to cells and which to background
  • Important: save your project frequently! Ilastik can crash!
  • Use Live Update to visualize training
  • Evaluate the result by checking Segmentation (Layer 1)
  • You can use the Uncertainty to see where more training points need to be added
  • Focus attention on pixels in between cells
  • Once the segmentation looks good, check a few other frames and update training as needed, until it looks good for all frames.
  • Go to Prediction Export
  • See details here
  • In Export Settings select Probabilities in the source field.
  • Open Choose Export Image Setting and select hdf5 format.
  • Then click Export All
  • Select as output folder ~/I2ICourse/Project2A/ProcessedData/

The output should be stored under the name [input_file_name]_Probabilities.h5, i.e. pos0_preproc-rg_Probabilities.h5.

Post-process with Python - Personal Computer

Follow these instruction if you run the python code locally (see below for instructions on how to run it on the cloud computer)

Launch Jupyter Labs

  • Navigate to the project folder, activate the conda environment, and launch Jupyter Labs:
cd ~/I2ICourse/
conda activate i2i_env
jupyter lab
  • In Jupyter Labs, navigate to spring_school_bioinformatics_microbiology/projects/project2/Project2A/
  • Then open the 0_post_process_segementation.ipynb notebook
  • Now run the notebook, see here for instructions on Jupyter Labs and follow the instructions in the notebook.

Post-process with Python - Cloud Computer

Follow these instructions if you are using the cloud computer to run Python.

Export Data on local computer

For the next steps we will switch to the cloud computers, but before that we need to transfer the data-file we just created:

  • On your local computer, compress the ~/I2ICourse/Project2A/ProcessedData folder into a zip-file
  • Upload this zip file to a cloud drive
  • Create a public share link and copy the address

Setup project folders on cloud computer

  • Login to the cloud computer (your tutor will provide info in how to do this)
  • Launch the command window and create a Project2A folder within the ~\workdir\ folder:
cd ~/workdir/
mkdir Project2A

Important Note: on the cloud computer we need to store all data in the ~/workdir/ folder or sub-folders of this, to make sure that files remain available after restarting the instance.

Also add a ProcessedData subfolder to the Project2A folder:

cd Project2A
mkdir ProcessedData

Download data on cloud computer

Now we download the data file we just uploaded to the cloud:

cd ~/workdir/Project2A/ProcessedData
curl -o  data.zip public_link_to_your_zip_file
unzip -j data.zip
  • Check that the data transferred successfully (use ls)
  • If so, then you can remove the zip file use rm data.zip

Download project code on cloud computer

  • Navigate to the workdir folder and use the git clone command to download the course code:
cd ~/workdir/
git clone https://github.com/sib-swiss/spring_school_bioinformatics_microbiology.git
  • This will create the folder ~/workdir/spring_school_bioinformatics_microbiology/ which contains all the Jupyter notebooks as well as the other course files

Launch Jupyter Labs

  • Next navigate to the project folder, activate the conda environment, and launch Jupyter Labs:
cd ~/workdir/
conda activate project2
jupyter lab
  • In Jupyter Labs, navigate to spring_school_bioinformatics_microbiology/projects/project2/Project2A/
  • Then open the 0_post_process_segementation.ipynb notebook
  • Now run the notebook, see here for instructions on Jupyter Labs and follow the instructions in the notebook.