Multiplexed images to AnnData#

SpatialTis could help you transform your segmented multiplexed images data into AnnData. You need to prepare:

  • Stacked .tiff-alike images

  • Cell mask images

Warning

SpatialTis CAN’T do segmentation, there are lots of resources out there that can help you with this part. You can try MCMICRO, ilastik, DeepCell and many more.

>>> import spatialtis as st

Specify your images and masks paths. You may also specify the marker name for each channel and add information of your images.

>>> images = ['image1.tif', 'image2.tif']
>>> masks = ['image1_mask.tif', 'image2_mask.tif']
>>> markers = pd.DataFrame(index=['CD20', 'CD8', 'CD99', 'NFkB'])
>>> annotations = pd.DataFrame({
...     "ROI": ["ROI1", "ROI2"],
...     "Tissue": ["Front", "Tail"],
... })

With all information ready, we can read it into AnnData. In the meantime, SpatialTis will extract the geometry information and expression matrix from the images.

>>> data = st.read_images(images, masks, markers=markers, annotations=annotations)
>>> data
    AnnData object with n_obs × n_vars = 2037 × 4
        obs: 'Tissue', 'ROI', 'area', 'cell_shape', 'centroid', 'eccentricity', ...
        obsm: 'spatial'