Skip to content

Segmentation and Tracking 2D Data with Delta2.0

Delta 2.0 is a Deep Learning based workflow that can segment and track 1D (mother machine) and 2D (family machine and agar pads) data. It uses two consecutive U-Net networks to first segment and then track cells. The pipeline is described in this publication. Detailed instruction can be found here

Note on data

The dataset we will work with here consists of a time-lapse data of a micro-colony of E. coli cells growing on LB agar pads. We have two channels: phase contrast and GFP. The GFP signal comes from a transcriptional reporter for Colicin Ib, a bacteriocin that is regulated by SOS-stress response. Images were taken every 5min.

More info on the data can be found here.

Install Delta (Windows only) Note for windows users

For Mac/Linux users Delta was installed when we created the i2i_env conda environment. However, Delta is not available from Conda for Windows. We have to install in manually, following these steps:

Download Delta (via git)

cd ~/I2ICourse/
git clone https://gitlab.com/dunloplab/delta.git

Download Delta (manually)

Download code here. Unzip file in ~/I2ICourse/ dir.

Add Delta path to each Notebook

Add the start of every notebook where we import delta, add the following lines at the top of the cell where we import delta:

import sys
sys.path.append(“path/to/delta”)

Getting started - Local Computer

  • On your computer, navigate to the I2ICourse folder and create a Project2B subfolder:
cd ~/I2ICourse/
mkdir Project2B
  • Then 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/Project2B/
  • Then follow the steps below.

Getting started - Cloud Computer

We will work on the cloud computers for the full project

Important: 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.

  • On the cloud computer, navigate to the workdir folder and create a Project2B subfolder:
cd ~/workdir/
mkdir Project2B
  • Then 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/Project2B/
  • Then follow the steps below.

Download test data

Run the 0_download_model_delta notebook to download the pre-trained network and data (note this may take some time, best to run this over a break!)

Run the pipeline

Work trough the 1_run_pipeline_delta notebook to see how the Delta2 pipeline works.

Note: processing speeds on CPU are very slow so use a CUDA compatible GPU based computer whenever possible.

Note: we also provided some instructions on how to use Delta on the Scicore cluster of University Basel [see scicore.md], to use Delta on clusters of other institutions, please contact your local cluster managers.

You can also try-out Delta on Google CoLabs.

Analyze the results

Work trough the 2_post_processing_delta notebook.

Analyze the data

Once you have the data analyzed, try to extract biological insight from it. Discuss with your tutor what question you could address. For this step on we encourage people to team-up in pairs/small-groups and work together. You can use the 3_explore_data_delta notebook as a starting point.

Note: use hardware acceleration on Apple Silicon

Create a Conda environment

conda create env -f environment_appleM1.yml

This was tested with MiniConda for Apple M1 installed from here, if you run into problems try uninstalling your current conda version and install the version above. The .env can be found in the project2 repository.

Git clone Delta

cd ~choose/a/path/for/delta  (below I use ~/miniconda/delta )
git clone https://gitlab.com/dunloplab/delta.git

Add Delta to search path (permanently)

add the following line to ~/.zshrc or ~/.bashrc file (depending on if you have a bash or zsh shell):

export PYTHONPATH="${PYTHONPATH}:~/miniconda/delta"

Change ~/miniconda/delta in the line above to the path you used in step 2a

After that you can run the notebooks without further changes.

You can check if you have access to GPU by running:

import tensorflow as tf
tf.config.list_physical_devices()