Whole Slide Example

Here we will show how to import OME-TIFF whole-slide images into CellTune.
The same folder organization applies to QPTIFF data as well. For more info see the data-preparation documentation

1. Create your project folder

Inside your CellTune_Store/CellTune_Data directory, create a folder for your project, for example:

CellTune_Store/CellTune_Data/Melanoma_Nov0725

Within this folder, create the following subfolders:

Images/
Segmentations/
Tables/

2. [Optional] Prepare your Tables

Create your tables:

Tables/MarkerTable.csv
Tables/CellTypeTable.csv

For instructions see documentation for MarkerTable(/documentation/getting-started/data-preparation/marker-table) and CellTypeTable(/documentation/getting-started/data-preparation/celltype-table) preparation.

You can also place any image/patient metadata tables here for organization. We plan that CellTune will utilize this data in the future.

3. Add your Whole-Slide Images

Organize all OME-TIFF images into the Images/ directory:

Images/WholeSlide1.ome.tiff
Images/WholeSlide2.ome.tiff
...

When prompted in CellTune, create a new project folder CellTune_Store/**CellTune_Projects**/Melanoma_Nov0725, and then select this Images directory as your image folder.

4. Add Segmentations (External)

Place the segmentation masks in the Segmentations/ folder using matching filenames with suffix ‘_segmentation_labels.tif’:

Segmentations/WholeSlide1_segmentation_labels.tif
Segmentations/WholeSlide2_segmentation_labels.tif
...

5. Zarr generation notes

Generating OME-Zarr from large whole-slide images may take a long time, depending on your system resources and drive read/write speed, and it also requires substantial memory. In our initial tests, the most efficient workflow was to copy the data locally, run the New Project Setup to generate the Zarrs, and then move or delete the temporary copy afterward.

Ensure you have sufficient RAM (>16 GB recommended) and enough local disk space (approximately 2× the size of the compressed TIFF or OME-TIFF/QPTIFF images) before generating the Zarrs.

We are actively working on improving generation speed and more effective use of remote/networked storage in future releases, as well as the option to skip Zarr generation entirely when data is already stored in an efficient format by the user and no editing is planned.

Tip: Before using CellTune, our lab generally prefers to pre-process the data by converting each OME-TIFF image into a single TIFF directory and performing background cleaning with Ilastik pixel classifiers. This often produces cleaner inputs, making it easier to classify, and can also reduce image size significantly.