SPACEc: Patch proximity analysis
Patch proximity analysis (PPA) analyses neighborhoods as patches of closely connected cells. The goal of the analysis is to analyze was surrounds these patches within a user defined radius. In our example we will use PPA to identify germinal centers as CN patches and then analyze was surrounds them based on the tissue condition (tonsil vs. tonsillitis).
# import spacec
import spacec as sp
#import standard packages
import os
import scanpy as sc
# silencing warnings
import warnings
warnings.filterwarnings('ignore')
sc.settings.set_figure_params(dpi=80, facecolor='white')
INFO:root: * TissUUmaps version: 3.1.1.6 /home/tim/miniforge3/envs/spacec/lib/python3.9/site-packages/numba/np/ufunc/parallel.py:371: NumbaWarning: The TBB threading layer requires TBB version 2021 update 6 or later i.e., TBB_INTERFACE_VERSION >= 12060. Found TBB_INTERFACE_VERSION = 12050. The TBB threading layer is disabled. warnings.warn(problem) /home/tim/miniforge3/envs/spacec/lib/python3.9/site-packages/cudf/utils/gpu_utils.py:89: UserWarning: A GPU with NVIDIA Volta™ (Compute Capability 7.0) or newer architecture is required. Detected GPU 0: NVIDIA GeForce GTX 1070 Detected Compute Capability: 6.1 warnings.warn(
# Specify the path to the data
root_path = "/home/user/path/SPACEc/" # inset your own path
data_path = root_path + 'example_data/raw/' # where the data is stored
# where you want to store the output
output_dir = root_path + 'example_data/output/'
os.makedirs(output_dir, exist_ok=True)
# read in the annotated anndata
adata = sc.read(output_dir + 'adata_nn_demo_annotated_cn.h5ad')
adata
AnnData object with n_obs × n_vars = 46789 × 59
obs: 'DAPI', 'x', 'y', 'area', 'region_num', 'unique_region', 'condition', 'leiden_1', 'leiden_1_subcluster', 'cell_type_coarse', 'cell_type_coarse_subcluster', 'cell_type_coarse_f', 'cell_type_coarse_f_subcluster', 'cell_type_coarse_f_f', 'cell_type', 'CN_k20_n40', 'CN_k20_n30', 'CN_k20_n20', 'CN_k20_n25', 'CN_k20_n6', 'CN_k20_n6_annot'
uns: 'CN_k20_n6_colors', 'Centroid_k20_n20', 'Centroid_k20_n25', 'Centroid_k20_n30', 'Centroid_k20_n40', 'Centroid_k20_n6', 'cell_type_coarse_f_colors', 'cell_type_colors', 'dendrogram_cell_type_coarse_f_subcluster', 'leiden', 'leiden_1_colors', 'neighbors', 'ppa_result_100', 'ppa_result_150', 'ppa_result_200', 'ppa_result_250', 'ppa_result_50', 'triDist', 'triDist_keyname', 'umap', 'unique_region_colors'
obsm: 'X_pca', 'X_umap'
layers: 'scaled'
obsp: 'connectivities', 'distances'
Setting the plotting parameter to True illustrates what the PPA function is detecting. This might be helpful if the min cluster size is unknown or the user wants to perform quality control. The results are stored as df in the adata.uns slot.
Compute for proximal cells
# this region result is also saved to adata.uns
region_results = sp.tl.patch_proximity_analysis(
adata,
region_column = "unique_region", # column with the region information
patch_column = "CN_k20_n6_annot", # column with the patch information (derive patches from this column)
group="Germinal Center", # group to consider
min_cluster_size=50, # minimum cluster size to consider
x_column='x', y_column='y', # spatial coordinates
radius = (5/0.5085), # to get the distance in µm
edge_neighbours = 3, # number of neighbours to consider for edge detection
key_name = 'ppa_result_5', # key name to store the result in adata.uns
plot = True) # plot detection for demonstration purposes
Estimated number of clusters: 4
Estimated number of noise points: 7
Finished reg002_Germinal Center
Estimated number of clusters: 2
Estimated number of noise points: 1
Finished reg001_Germinal Center
region_results
( DAPI x y area region_num unique_region \
0 121.210000 2176.510000 1814.550000 100.0 1 reg002
2 117.117021 1995.148936 1462.212766 94.0 1 reg002
0 129.582524 2429.368932 1558.776699 103.0 0 reg001
condition leiden_1 leiden_1_subcluster cell_type_coarse ... \
0 tonsillitis 6 6,0 GCB ...
2 tonsillitis 15 15,1 B cell ...
0 tonsil 15 15,1 B cell ...
cell_type_coarse_f_f cell_type CN_k20_n40 CN_k20_n30 CN_k20_n20 CN_k20_n25 \
0 GCB GCB 1 1 1 1
2 B cell B cell 1 1 1 1
0 B cell B cell 1 1 1 1
CN_k20_n6 CN_k20_n6_annot patch_id unique_patch_ID
0 1 Marginal Zone 2 reg002_Marginal Zone_patch_no_2
2 1 Marginal Zone 2 reg002_Marginal Zone_patch_no_2
0 1 Marginal Zone 1 reg001_Marginal Zone_patch_no_1
[3 rows x 23 columns],
DAPI x y area region_num unique_region \
1837 113.646724 396.188034 568.669516 351.0 1 reg002
1745 63.307339 382.293578 560.876147 218.0 1 reg002
1751 49.727679 383.803571 581.084821 224.0 1 reg002
1972 74.031519 413.765043 581.971347 349.0 1 reg002
2004 54.897098 419.155673 599.828496 379.0 1 reg002
... ... ... ... ... ... ...
46358 49.146409 2494.881215 1466.886740 362.0 0 reg001
46285 58.472813 2485.087470 1501.004728 423.0 0 reg001
46525 50.608407 2512.601770 1455.918142 452.0 0 reg001
46358 49.146409 2494.881215 1466.886740 362.0 0 reg001
46648 84.843949 2525.611465 1475.210191 314.0 0 reg001
condition leiden_1 leiden_1_subcluster cell_type_coarse ... \
1837 tonsillitis 0 0 B cell ...
1745 tonsillitis 6 6,1 GCB ...
1751 tonsillitis 6 6,1 GCB ...
1972 tonsillitis 6 6,0 GCB ...
2004 tonsillitis 6 6,0 GCB ...
... ... ... ... ... ...
46358 tonsil 6 6,3 GCB ...
46285 tonsil 6 6,3 GCB ...
46525 tonsil 6 6,3 GCB ...
46358 tonsil 6 6,3 GCB ...
46648 tonsil 6 6,3 GCB ...
cell_type_coarse_f_subcluster cell_type_coarse_f_f cell_type CN_k20_n40 \
1837 B cell B cell B cell 4
1745 GCB GCB GCB 4
1751 GCB GCB GCB 4
1972 GCB GCB GCB 4
2004 GCB GCB GCB 4
... ... ... ... ...
46358 GCB GCB GCB 4
46285 GCB GCB GCB 4
46525 GCB GCB GCB 4
46358 GCB GCB GCB 4
46648 GCB GCB GCB 4
CN_k20_n30 CN_k20_n20 CN_k20_n25 CN_k20_n6 CN_k20_n6_annot cluster
1837 4 4 4 4 Germinal Center 0
1745 4 4 4 4 Germinal Center 0
1751 4 4 4 4 Germinal Center 0
1972 4 4 4 4 Germinal Center 0
2004 4 4 4 4 Germinal Center 0
... ... ... ... ... ... ...
46358 4 4 4 4 Germinal Center 1
46285 4 4 4 4 Germinal Center 1
46525 4 4 4 4 Germinal Center 1
46358 4 4 4 4 Germinal Center 1
46648 4 4 4 4 Germinal Center 1
[939 rows x 22 columns])
Often it is more informative to derive the cellular content within a range of distances.
pixel_to_um = 0.5085 # conversion factor from pixel to µm
distances = [5/pixel_to_um, 10/pixel_to_um, 15/pixel_to_um, 20/pixel_to_um, 25/pixel_to_um] # distances in µm
key_names = ['ppa_result_5', 'ppa_result_10', 'ppa_result_15', 'ppa_result_20', 'ppa_result_25'] # key names to store the results
for distance, key_name in zip(distances, key_names): # loop over distances and key names
# this region result is also saved to adata.uns
region_results = sp.tl.patch_proximity_analysis(
adata,
region_column = "unique_region",
patch_column = "CN_k20_n6_annot",
group="Germinal Center",
min_cluster_size=50,
x_column='x', y_column='y',
radius = distance, # to get the distance in µm
edge_neighbours = 3,
key_name = key_name,
plot = False)
Estimated number of clusters: 4
Estimated number of noise points: 7
Finished reg002_Germinal Center
Estimated number of clusters: 2
Estimated number of noise points: 1
Finished reg001_Germinal Center
Estimated number of clusters: 4
Estimated number of noise points: 7
Finished reg002_Germinal Center
Estimated number of clusters: 2
Estimated number of noise points: 1
Finished reg001_Germinal Center
Estimated number of clusters: 4
Estimated number of noise points: 7
Finished reg002_Germinal Center
Estimated number of clusters: 2
Estimated number of noise points: 1
Finished reg001_Germinal Center
Estimated number of clusters: 4
Estimated number of noise points: 7
Finished reg002_Germinal Center
Estimated number of clusters: 2
Estimated number of noise points: 1
Finished reg001_Germinal Center
Estimated number of clusters: 4
Estimated number of noise points: 7
Finished reg002_Germinal Center
Estimated number of clusters: 2
Estimated number of noise points: 1
Finished reg001_Germinal Center
# save adata
adata.write(output_dir + 'adata_nn_demo_annotated_cn.h5ad')
... storing 'region_num' as categorical
... storing 'unique_region' as categorical
... storing 'condition' as categorical
... storing 'leiden_1' as categorical
... storing 'leiden_1_subcluster' as categorical
... storing 'cell_type_coarse' as categorical
... storing 'cell_type_coarse_subcluster' as categorical
... storing 'cell_type_coarse_f' as categorical
... storing 'cell_type_coarse_f_subcluster' as categorical
... storing 'cell_type_coarse_f_f' as categorical
... storing 'cell_type' as categorical
... storing 'CN_k20_n6_annot' as categorical
SPACEc can visualize the PPA results as donut plot showing the percentages of cell types or CNs within a given radius around the patches. Percentages are averaged over all regions in the selected condition. The donut plot can show up to five distances.
Visualization
# Donut plots for cell types around Germinal Center
sp.pl.ppa_res_donut(adata,
palette=None, # color palette
cat_col = "cell_type", # column with the cell type or CN information
key_names = ['ppa_result_5', 'ppa_result_10', 'ppa_result_15', 'ppa_result_20', 'ppa_result_25'], # key names to plot
radii = [5, 10, 15, 20, 25], # radii in µm
unit = 'µm', # unit
figsize = (10,10), # figure size
add_guides = True, # add guides
text = 'Cell types around Germinal Center', # text to display in the middle
label_color='white', # distance label color
subset_column = 'condition', # column to subset
subset_condition = 'tonsil', # condition to subset
title='Tonsil patch proximity analysis') # title
sp.pl.ppa_res_donut(adata,
palette=None,
cat_col = "cell_type",
key_names = ['ppa_result_5', 'ppa_result_10', 'ppa_result_15', 'ppa_result_20', 'ppa_result_25'],
radii = [5, 10, 15, 20, 25],
unit = 'µm',
figsize = (10,10),
add_guides = True,
text = 'Cell types around Germinal Center',
label_color='white',
subset_column = 'condition',
subset_condition = 'tonsillitis',
title='Tonsillitis patch proximity analysis')
Key 0: ppa_result_25
Key 0 has 1720 rows.
Key 1: ppa_result_20
Key 1 has 1271 rows.
Key 2: ppa_result_15
Key 2 has 850 rows.
Key 3: ppa_result_10
Key 3 has 456 rows.
Key 4: ppa_result_5
Key 4 has 3 rows.
Key 0: ppa_result_25
Key 0 has 1720 rows.
Key 1: ppa_result_20
Key 1 has 1271 rows.
Key 2: ppa_result_15
Key 2 has 850 rows.
Key 3: ppa_result_10
Key 3 has 456 rows.
Key 4: ppa_result_5
Key 4 has 3 rows.
# Donut plots for CNs around Germinal Center
sp.pl.ppa_res_donut(adata,
palette=None,
cat_col = "CN_k20_n6_annot",
key_names = ['ppa_result_5', 'ppa_result_10', 'ppa_result_15', 'ppa_result_20', 'ppa_result_25'],
radii = [5, 10, 15, 20, 25],
unit = 'µm',
figsize = (10,10),
add_guides = True,
text = 'CNs around Germinal Center',
label_color='white',
subset_column = 'condition',
subset_condition = 'tonsil',
title='Tonsil patch proximity analysis')
sp.pl.ppa_res_donut(adata,
palette=None,
cat_col = "CN_k20_n6_annot",
key_names = ['ppa_result_5', 'ppa_result_10', 'ppa_result_15', 'ppa_result_20', 'ppa_result_25'],
radii = [5, 10, 15, 20, 25],
unit = 'µm',
figsize = (10,10),
add_guides = True,
text = 'CNs around Germinal Center',
label_color='white',
subset_column = 'condition',
subset_condition = 'tonsillitis',
title='Tonsillitis patch proximity analysis')
Key 0: ppa_result_25
Key 0 has 1720 rows.
Key 1: ppa_result_20
Key 1 has 1271 rows.
Key 2: ppa_result_15
Key 2 has 850 rows.
Key 3: ppa_result_10
Key 3 has 456 rows.
Key 4: ppa_result_5
Key 4 has 3 rows.
Key 0: ppa_result_25
Key 0 has 1720 rows.
Key 1: ppa_result_20
Key 1 has 1271 rows.
Key 2: ppa_result_15
Key 2 has 850 rows.
Key 3: ppa_result_10
Key 3 has 456 rows.
Key 4: ppa_result_5
Key 4 has 3 rows.
