Tutorial : 10x Visium(Breast Cancer)

We utilized spatial transcriptomics data from the 10x Visium platform focusing on breast cancer. The dataset originates from the study Spatial Transcriptomics Reveal Pitfalls and Opportunities for the Detection of Rare High-Plasticity Breast Cancer Subtypes.

import os
os.environ['R_HOME']="/share/home/baixiaosheng/software/anaconda/envs/SpaGT/lib/R"
import pandas as pd
import numpy as np
import scanpy as sc
import anndata as ad
import random, torch
import warnings
warnings.filterwarnings("ignore")
import matplotlib.pyplot as plt
from sklearn.metrics.cluster import adjusted_rand_score
import SpaGT
SpaGT.__version__

' 1.0.0 '

file = './cl.h5ad'
adata= ad.read_h5ad(file)

adata, conns = SpaGT.spatial_reconstruction(adata,alpha=1.0)

features_tensor = torch.tensor(adata.obsm['expression_embedding']).float()
relations_tensor = torch.tensor(conns).float()

SpaGT.setup_seed(303)

n_clusters = 5
radius = 50

clf=SpaGT.DEC(features_tensor,relations_tensor,node_width=features_tensor.shape[1], edge_width=relations_tensor.shape[0],num_heads=4,nhid=28,n_clusters=n_clusters)


#Run
a = clf.fit(features_tensor,relations_tensor, init="leiden",num_cluster=n_clusters,res=0.2,n_neighbors=20,pca_num=23)
y_pred, prob,z1=clf.predict()


adata.obs["pred"] = y_pred
adata.obs["pred"] = adata.obs["pred"].astype('category')

leiden_n_clusters=len(np.unique(y_pred))
print("class:",leiden_n_clusters)

class: 5

adata, refine = SpaGT.refine_label(adata, radius=radius,key='pred')

purity = SpaGT.purity_score(adata.obs['annotation'], adata.obs['pred'])
ARI = adjusted_rand_score(adata.obs['annotation'], adata.obs['pred'])
print("purity: ", purity)
print("ARI: ", ARI)

purity:  0.7171052631578947
ARI:  0.2964973818413119

refine_purity = SpaGT.purity_score(adata.obs['annotation'], adata.obs['label_refined'])
refine_ARI = adjusted_rand_score(adata.obs['annotation'], adata.obs['label_refined'])
print("refine_purity: ", refine_purity)
print("refine_ARI: ", refine_ARI)

refine_purity:  0.6626794258373205
refine_ARI:  0.26464866849348045

plot_color=["#70a1d7","#ffb400","#f7f48b","#59BE86","#5fc9f8","#F56867","#dedad0"]
sc.pl.spatial(
    adata,
    img_key=None,
    color='pred',
    size=1.5,
    palette=plot_color,
    legend_loc='right margin',
    show=False,
    title=[f'SpaGT \n(purity={purity:.3f} ARI={ARI:.3f})']
)

sc.pl.spatial(
    adata,
    img_key=None,
    color='label_refined',
    size=1.5,
    palette=plot_color,
    legend_loc='right margin',
    show=False,
    title=[f'SpaGT \n(refine_purity={refine_purity:.3f} ARI={refine_ARI:.3f})']
)

plt.show()