Skip to content

Example Data

Jump to bottom
Nils edited this page Jun 25, 2021 · 3 revisions

Example Data

This page will give a short overview on how to download and run the pipeline on example data.

Download

You can download free example data via this release page: https://github.com/RUB-Bioinf/HT-PropagatedNeuriteSkeletonization/releases/tag/0.0.9

This release provides two sets of different phenotypic cell outgrowths and stainings. In the published works, these sets were used during the statistical evaluations.

Note that you should still use the latest release for soruce code or Docker images!

Setup

Please follow the guide first on how to set up the docker image and run the pipeline. This section assumes you are familiar with Docker and have registered an input path (refereed to as resources from here on out, as that is the same name as it will be linked as from within the Docker) and output path (called output, for the same reasons).

Once you have downloaded the example data, follow this page on how to run the pipeline on the example data:

Resources

Use this directory for all your input data. Make sure you are using 16-bit tif image files. You can place your files in different sub-directories. Every directory in the resources folder will be examined and scanned for images to process.

We named our cell channels based on the antibody and techniques used in the presented works. As such, a pair of images would be called:

  • PK4_20200311_MDN_Plinks_150_01_Alexa488_01.tif
  • PK4_20200311_MDN_Plinks_150_01_DAPI_01.tif

or

  • PK4_20200311_MDN_ctrllinks_0_01_Alexa488_01.tif
  • PK4_20200311_MDN_ctrllinks_0_01_DAPI_01.tif

So, as this example shows, the pipeline takes an input image (make sure it has sufficient filename character count!) and matches the Alexa488 channel to the DAPI channel. This way, both channels can be overlapped during runtime and morphological information extracted.

We named the channels after staining, as such, the DAPI file should only contain a grayscale image of your nuclei staining. The Alexa488 file a grayscale of your neurite staining.

As an example, this would be a valid file structure the pipeline can work with:

/ resources
  / A01
     / PK4_20200311_MDN_Plinks_150_01_Alexa488_01.tif
     / PK4_20200311_MDN_Plinks_150_01_DAPI_01.tif

     / PK4_20200311_MDN_Plinks_150_01_Alexa488_02.tif
     / PK4_20200311_MDN_Plinks_150_01_DAPI_02.tif

     / PK4_20200311_MDN_Plinks_150_01_Alexa488_03.tif
     / PK4_20200311_MDN_Plinks_150_01_DAPI_03.tif

     / ...

  / A02
     / PK5_20200311_MDN_Plinks_150_01_Alexa488_01.tif
     / PK5_20200311_MDN_Plinks_150_01_DAPI_01.tif

     / ...

In practice, this example would be used, if you were to use plated in-vitro cells and wanted to organize them in folders named after their origin wells.

If you have downloaded the provided example data (see above) you will see that the naming conventions and data structures have changed. This is no problem and will work for evaluations as well.

Using the provided data, the structure should look like this:

/ resources
  / primary-mesencephalic-dopaminergic-cells
     / ...

  / example-SH-SY5Y
     / ...

Output

This is the output directory for everything the pipeline calculates. During the calculations, multiple files will be created and overwritten here. Whenever you run the pipeline, a new directory is created storing all results of the current run.

The final results of the pipeline will contain images of the processed data. You can use them as vaildation to check if your parameters are correct. You will also find various CSV files of your statistics that you can import into Microsoft Excel for further evaluations.

Clone this wiki locally