%matplotlib inline

Extract summary features

This example shows how to extract summary features from the tissue image.

Summary features give a good overview over the intensity of each image channels at the location of the Visium spots. They are calculated by using features = 'summary', which calls squidpy.im.ImageContainer.features_summary().

In addition to feature_name and channels we can specify the following features_kwargs:

• quantiles - quantiles that are computed. By default, the 0.9th, 0.5th, and 0.1th quantiles are calculated.

See also

See {doc}`compute_features` for general usage
of {func}`squidpy.im.calculate_image_features`.

import squidpy as sq

First, let’s load the fluorescence Visium dataset.

# get spatial dataset including hires tissue image
img = sq.datasets.visium_fluo_image_crop()
adata = sq.datasets.visium_fluo_adata_crop()

Then, we calculate the 0.1th quantile, mean and standard deviation for the Visium spots of the fluorescence channels 0 (DAPI) and 1 (GFAP). In order to get statistics of only the tissue underneath the spots, we use the argument mask_circle = True. When not setting this flag, statistics are calculated using a square crop centered on the spot.

# calculate summary features and save in key "summary_features"
sq.im.calculate_image_features(
    adata,
    img,
    features="summary",
    features_kwargs={
        "summary": {
            "quantiles": [0.1],
            "channels": [0, 1],
        }
    },
    key_added="summary_features",
    mask_circle=True,
    show_progress_bar=False,
)

The result is stored in adata.obsm['summary_features'].

adata.obsm["summary_features"].head()

	summary_ch-0_quantile-0.1	summary_ch-0_mean	summary_ch-0_std	summary_ch-1_quantile-0.1	summary_ch-1_mean	summary_ch-1_std
AAACGAGACGGTTGAT-1	0.0	6849.760120	14383.136990	0.0	3762.582691	2462.922155
AAAGGGATGTAGCAAG-1	0.0	4469.448519	11304.932832	0.0	3824.862145	2153.804234
AAATGGCATGTCTTGT-1	0.0	5944.567897	9808.327041	0.0	5481.824787	6747.728831
AAATGGTCAATGTGCC-1	0.0	5259.799257	9115.113451	0.0	2628.194501	1418.504292
AAATTAACGGGTAGCT-1	0.0	4468.428701	10285.605481	0.0	4036.154302	4447.304626

Use squidpy.pl.extract to plot the summary features on the tissue image or have a look at our interactive visualization tutorial to learn how to use our interactive napari plugin. Note how the spatial distribution of channel means is different for fluorescence channels 0 (DAPI stain) and 1 (GFAP stain).

sq.pl.spatial_scatter(
    sq.pl.extract(adata, "summary_features"),
    color=[None, "summary_ch-0_mean", "summary_ch-1_mean"],
    img_cmap="gray",
)