💡Simple Pipeline

This step-by-step tutorial is meant to implement the most basic pipeline for using both QuPath and Python to process Image Detection.

The Getting Set Up page explains what and how to download everything needed in this tutorial.

Pipeline of the Project. As a first step, all the H-DAB stained slides need to be slide scanned and stored in a Folder. In QuPath a project will be created as a final step and Object classifier gets trained recognising detected objects as positive cells, negative cells, or no cell artifacts. With git bash a prewritten Python script gets called resulting in an outcome sheet linking the ID_Slidescanning with the ID_Sample and providing us with the positivity rate of each sample. With the merger, the donor information gets linked with the positivity rate resulting in a final Excel spreadsheet.

The prewritten QuPath script that will be downloaded at the end of the tutorial needs information that's specific to your slides. Therefore we start by creating a project in QuPath and creating an object classifier, that distinguishes between immuno-positive and negative cells. To show the script where it needs to detect and classify the cells either a manual region detection or an automatic threshold can be chosen.

Afterward, the project-specific parts get integrated into the code and we can run a batch analysis on hundreds of slides. The results get processed by a Python script that you can locally run, resulting in a spreadsheet containing the positivity rate for each slide.

Step 1: Opening Scans in QuPath

Step 2: Train object classifier

Step 3: Running QuPath script locally

Step 4: Collect the results file

Step 5: Move the Inventory file to the same directory

Step 6: Call ProcessScan Class in the directory

Step 7: Create a final spreadsheet merging everything