脚本更新---高精度空转（Xenium、CosMx）Patch Proximity Analysis

原创

追风少年i

发布于 2025-12-09 10:17:15

1720

作者，Evil Genius

周一了，小伙伴们，我们来更新一下脚本吧。

这个分析就是依据邻域范围分析空间结构

我们来分享一下

# 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')
# 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

邻近细胞计算

这个还是之前提到的两种策略：第一种方法“border_cell_radius”会识别勾勒凹包轮廓的细胞及其最近邻细胞，随后以这些细胞为中心、按指定半径绘制圆形区域，以检测空间上邻近的细胞，同时忽略属于同一斑块的细胞。此方法能生成高度精确的边界，但计算强度通常高于“hull_expansion”方法。此外，“border_cell_radius”方法对内部边界高度敏感：若半径范围覆盖斑块内部孔洞，该方法可有效识别这些孔洞结构。

# this region result is also saved to adata.uns
results, outlines_results = sp.tl.patch_proximity_analysis(
    adata, # the annotated adata object
    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 = [20, 40, 60, 80, 100], # to get the distance in µm
    edge_neighbours = 1, # number of neighbours to consider for edge detection if set to 1 only the hull is considered
    plot = True, # plot the results for demonstration and/or documentation (set to False to skip plotting - improves speed)
    original_unit_scale = 1.96656, # scale factor for the units (1 = 1px per unit e.g. µm)
    method= "border_cell_radius", # method to use for the edge detection
    key_name = "ppa_result_20_40_60_80_100_border_cell_radius", # key name to store the result in adata.uns
    save_geojson = False, # save the results as geojson
    ) # plot detection for demonstration purposes

##Processing reg002_Germinal Center
##Estimated number of clusters: 4
##Estimated number of noise points: 7

results

可视化

# Donut plots for cell types around Germinal Center
sp.pl.ppa_res_donut(
    adata,
    cat_col = 'cell_type',
    key_name="ppa_result_20_40_60_80_100_border_cell_radius",
    palette=None,
    distance_mode="within",  # "within" or "between"
    unit="µm",
    figsize=(10, 10),
    add_guides=True,
    text="Cell types around Germinal Center",
    label_color="black",
    group_by= 'condition',
    title="PPA",
)

# Donut plots for cell types around Germinal Center
sp.pl.ppa_res_donut(
    adata,
    cat_col = 'cell_type',
    key_name="ppa_result_20_40_60_80_100_border_cell_radius",
    palette=None,
    distance_mode="between",  # "within" or "between"
    unit="µm",
    figsize=(10, 10),
    add_guides=True,
    text="Cell types around Germinal Center",
    label_color="black",
    group_by= 'condition',
    title="PPA",
)

# Donut plots for cell types around Germinal Center
sp.pl.ppa_res_donut(
    adata,
    cat_col = 'CN_k20_n6_annot',
    key_name="ppa_result_20_40_60_80_100_border_cell_radius",
    palette=None,
    distance_mode="within",  # "within" or "between"
    unit="µm",
    figsize=(10, 10),
    add_guides=True,
    text="Cell types around Germinal Center",
    label_color="black",
    group_by= 'condition',
    title="PPA",
)