{
"cells": [
{
"cell_type": "markdown",
"id": "9f8f826b-acab-4ce2-a04d-c44ef3d97223",
"metadata": {},
"source": [
"# Downstream Analysis\n",
"\n",
"After performing all of the preprocessing with `spatialproteomics`, you typically want to use other packages to extract further information from the data. Here, we show two examples, one using `squidpy` and the other one using the R package `spatstat`. `Spatialproteomics` provides several ways to export your data to facilitate interoperability with external tools.\n",
"\n",
"If you want to follow along with this tutorial, you can download the data [here](https://oc.embl.de/index.php/s/XzEa9po1tjiDzzJ)."
]
},
{
"cell_type": "code",
"execution_count": 1,
"id": "cfa21d6e-492c-4e7f-9449-81c2ef3236c9",
"metadata": {
"tags": []
},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/home/meyerben/meyerben/.conda/envs/tmp_env_3/lib/python3.10/site-packages/dask/dataframe/__init__.py:31: FutureWarning: The legacy Dask DataFrame implementation is deprecated and will be removed in a future version. Set the configuration option `dataframe.query-planning` to `True` or None to enable the new Dask Dataframe implementation and silence this warning.\n",
" warnings.warn(\n"
]
}
],
"source": [
"import os\n",
"import sys\n",
"\n",
"# setup for rpy2 to use the correct R binary\n",
"env_r_bin = \"/g/huber/users/meyerben/.conda/envs/tmp_env_3/bin\"\n",
"os.environ[\"PATH\"] = env_r_bin + os.pathsep + os.environ.get(\"PATH\", \"\")\n",
"if \"R_HOME\" in os.environ:\n",
" del os.environ[\"R_HOME\"]\n",
"\n",
"%load_ext rpy2.ipython\n",
"\n",
"import xarray as xr\n",
"import spatialproteomics as sp\n",
"import pandas as pd\n",
"import squidpy as sq\n",
"\n",
"celltype_colors = {\n",
" \"B cell\": \"#5799d1\",\n",
" \"T cell\": \"#ebc850\",\n",
" \"Myeloid cell\": \"#de6866\",\n",
" \"Dendritic cell\": \"#4cbcbd\",\n",
" \"Macrophage\": \"#bb7cb4\",\n",
" \"Stromal cell\": \"#62b346\",\n",
" \"Endothelial cell\": \"#bf997d\",\n",
"}"
]
},
{
"cell_type": "code",
"execution_count": 2,
"id": "f998a188-d183-4add-a907-77f004ed987c",
"metadata": {
"tags": []
},
"outputs": [],
"source": [
"# loading in a data set\n",
"ds = xr.open_zarr(\"../../data/LN_24_1.zarr\")\n",
"# for clearer visualizations, we set the cell types to the broadest level (B cells, T cells, ...)\n",
"ds = ds.la.set_label_level(\"labels_0\", ignore_neighborhoods=True).la.set_label_colors(\n",
" celltype_colors.keys(), celltype_colors.values()\n",
")"
]
},
{
"cell_type": "markdown",
"id": "d99bcbbb-29ad-4f72-addd-7883f7988771",
"metadata": {},
"source": [
"## Neighborhood Enrichment with Squidpy\n",
"\n",
"We can use the `pp.convert_to_anndata()` function to turn the `spatialproteomics` object into an `anndata` object that is accepted by `squidpy`. We can then follow the [squidpy documentation](https://squidpy.readthedocs.io) to perform an enrichment analysis."
]
},
{
"cell_type": "code",
"execution_count": 3,
"id": "8099175e-20c1-497d-9666-60e3ac05266e",
"metadata": {
"tags": []
},
"outputs": [
{
"data": {
"text/plain": [
"AnnData object with n_obs × n_vars = 18134 × 56\n",
" obs: 'BCL-2_binarized', 'BCL-6_binarized', 'CD163_binarized', 'CD45RA_binarized', 'CD45RO_binarized', 'CD8_binarized', 'DAPI_binarized', 'FOXP3_binarized', 'Helios_binarized', 'ICOS_binarized', 'PD-1_binarized', 'TCF7-TCF1_binarized', 'Tim3_binarized', '_labels', '_neighborhoods', 'centroid-0', 'centroid-1', 'degree', 'diversity_index', 'homophily', 'inter_label_connectivity'\n",
" uns: '_labels_colors'\n",
" obsm: 'spatial'"
]
},
"execution_count": 3,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"adata = ds.tl.convert_to_anndata()\n",
"adata"
]
},
{
"cell_type": "code",
"execution_count": 4,
"id": "d43abb55-6856-4463-9fd2-9dcb57278da4",
"metadata": {
"tags": []
},
"outputs": [
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "fd9925bbba5e48f2979a3e1187f15663",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
" 0%| | 0/1000 [00:00"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"sq.gr.spatial_neighbors(adata, coord_type=\"generic\", radius=140)\n",
"sq.gr.nhood_enrichment(adata, cluster_key=\"_labels\")\n",
"sq.pl.nhood_enrichment(adata, cluster_key=\"_labels\")"
]
},
{
"cell_type": "markdown",
"id": "2a2872c0-dab1-4df3-a194-43fa735a07fb",
"metadata": {},
"source": [
"We see that in this sample, B cells frequently cluster together, while B and T cells have a negative enrichment score, meaning that the two do not mix very often."
]
},
{
"cell_type": "markdown",
"id": "cfb3a86e-14df-4fda-9ee7-3b4a6d47c1be",
"metadata": {},
"source": [
"## Downstream Analysis in R using spatstat\n",
"\n",
"R has a great analysis ecosystem for spatial data, hence it can be desirable to continue analysis with packages such as `spatstat`. Below we show how you can export your data in an R-friendly format and perform some basic analysis using the `spatstat package`."
]
},
{
"cell_type": "code",
"execution_count": 5,
"id": "10b29a9f-c1ff-4c7c-b61e-357a4d9f28cd",
"metadata": {
"tags": []
},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" centroid.0 | \n",
" centroid.1 | \n",
" labels | \n",
"
\n",
" \n",
" \n",
" \n",
" | 1 | \n",
" 343.043825 | \n",
" 1492.916335 | \n",
" Endothelial cell | \n",
"
\n",
" \n",
" | 2 | \n",
" 341.740506 | \n",
" 1517.082278 | \n",
" B cell | \n",
"
\n",
" \n",
" | 3 | \n",
" 342.250000 | \n",
" 1450.492857 | \n",
" B cell | \n",
"
\n",
" \n",
" | 4 | \n",
" 345.150602 | \n",
" 1559.259036 | \n",
" B cell | \n",
"
\n",
" \n",
" | 5 | \n",
" 346.454128 | \n",
" 1575.885321 | \n",
" T cell | \n",
"
\n",
" \n",
"
\n",
"