Tutorial 2: Data pre-processing for STARmap_PLUS data

STARmap PLUS enables high-resolution spatial transcriptomics along with protein detection in the same tissue section. The pre-processings aim to regist the extracellular proteins with cells.

import os
import scanpy as sc
import numpy as np
import pandas as pd

import matplotlib.pyplot as plt

Define the path and hyperparameters

N = 3
threshold_tau = 15
threshold_plaque = 50

name = '13months-disease-replicate_1_random.h5ad'
path = os.path.join('/home/wzk/ST_data/AD_mouse', name)
save_dir = '/home/wzk/ST_data/AD_mouse/'

rna_adata = sc.read_h5ad(path)
path_plaque = '/home/wzk/ST_data/AD_mouse/tau&abeta/plaque_13months-disease-replicate_1.csv'
tao_file = np.load('/home/wzk/ST_data/AD_mouse/tau&abeta/13months-disease-replicate_1_tau_center.npy')

plaque_file = pd.read_csv(path_plaque)

Plot RNA and MSI data in the original coodinates

fig, axs = plt.subplots(1, 3, figsize=(20, 6))

sc.pl.embedding(rna_adata, basis='spatial', color='leiden', title=f'Expression of rna (scanpy)', ax=axs[0], show=False)

x_max = tao_file[:, 0].max()

axs[1].set_title('Expression of rna (matplotlib)')
axs[1].scatter(rna_adata.obsm['spatial'][:, 0], rna_adata.obsm['spatial'][:, 1], color='red', s=0.5)  # 使用蓝色标记点

axs[2].set_title('Expression of plaque and tao (matplotlib)')
axs[2].scatter(plaque_file['m.cx'].values, plaque_file['m.cy'].values, color='black', s=5)  # 使用蓝色标记点
axs[2].scatter(rna_adata.obsm['spatial'][:, 0], rna_adata.obsm['spatial'][:, 1], color='black', s=0.2, )  # 使用蓝色标记点

if name[0] == '8':
    axs[2].scatter(x_max - tao_file[:, 0], tao_file[:, 1], color='green', s=0.5)
else:
    axs[2].scatter(tao_file[:, 0], tao_file[:, 1], color='green', s=0.5)

plt.show()

/home/wzk/anaconda3/envs/pytorch_zk/lib/python3.8/site-packages/scanpy/plotting/_tools/scatterplots.py:394: UserWarning: No data for colormapping provided via 'c'. Parameters 'cmap' will be ignored
  cax = scatter(

plaque_coordinates = np.array((plaque_file['m.cx'].values, plaque_file['m.cy'].values)).T
tao_coordinates = np.array((tao_file[:, 0], tao_file[:, 1])).T
cell_coordinates = rna_adata.obsm['spatial']

Calculate the distance matrix

distance_matrix_plaque_cell = np.sqrt(((plaque_coordinates[:, np.newaxis, :] - cell_coordinates[np.newaxis, :, :]) ** 2).sum(axis=2))
distance_matrix_tao_cell = np.sqrt(((tao_coordinates[:, np.newaxis, :] - cell_coordinates[np.newaxis, :, :]) ** 2).sum(axis=2))

temp_tao = np.argsort(distance_matrix_tao_cell, axis=1)[:, :N]#.reshape(-1)
temp_plaque = np.argsort(distance_matrix_plaque_cell, axis=1)[:, :N]#.reshape(-1)

Align proteins to cells according to the predefined distances

mask_tao = (distance_matrix_tao_cell.min(axis=1) < threshold_tau)
mask_plaque = (distance_matrix_plaque_cell.min(axis=1) < threshold_plaque)

temp_tao = temp_tao[mask_tao, ].reshape(-1)
temp_plaque = temp_plaque[mask_plaque, ].reshape(-1)

index_tao = np.zeros(rna_adata.shape[0])
for i in temp_tao.tolist():
    index_tao[i] = 1
print(index_tao.sum())

index_plaque = np.zeros(rna_adata.shape[0])
for i in temp_plaque.tolist():
    index_plaque[i] = 1
print(index_plaque.sum())

1116.0
429.0

rna_adata.obs['tao'] = index_tao
rna_adata.obs['plaque'] = index_plaque

Visualize the aligned data

fig, axs = plt.subplots(1, 2, figsize=(14, 6))

sc.pl.embedding(rna_adata, basis='spatial', color='tao', title=f'Expression of rna (scanpy)', ax=axs[0], show=False)
# sc.pl.embedding(rna_adata, basis='spatial', color='plaque', title=f'Expression of rna (scanpy)', ax=axs[1], show=False)

mask_tao = rna_adata.obs['tao'].values == 1
mask_plaque = rna_adata.obs['plaque'].values == 1
axs[1].scatter(rna_adata.obsm['spatial'][mask_tao, 0], rna_adata.obsm['spatial'][mask_tao, 1], color='green', s=0.5)
axs[1].scatter(rna_adata.obsm['spatial'][mask_plaque, 0], rna_adata.obsm['spatial'][mask_plaque, 1], color='black', s=3)

<matplotlib.collections.PathCollection at 0x7f622f4dcf70>