Mouse Brain Clustering Analysis

3:MouseBrain_Clustering¶

HiSTaR can process 10X Visium data to extract low-dimension representation. In this tutorial, we use Mouse Brain dataset to introduce the analysis.
The raw Mouse Brain dataset can be downloaded from: https://www.10xgenomics.com/resources/datasets/adult-mouse-brain-section-1-coronal-stains-dapi-anti-neu-n-1-standard-1-1-0
The complete experimental dataset is available here https://zenodo.org/records/15599070

Preparation¶

In [1]:

import scanpy as sc
import matplotlib.pyplot as plt
import torch
import warnings
warnings.filterwarnings("ignore")

In [2]:

import HiSTaR
HiSTaR.fix_seed(2023)

In [3]:

import matplotlib as mpl
mpl.rcParams.update({
    'font.family': 'Arial',             
    'axes.labelweight': 'bold',         
    'axes.titleweight': 'bold',         
    'axes.titlesize': 12,               
    'axes.titlelocation': 'left',       
    'figure.constrained_layout.use': True,  
    'figure.dpi': 300,                   
    'savefig.dpi': 300,
})

mclust_palette = [
    "#F3766E", "#5BB300", "#2E96FF", "#C655D9", "#FFB549",
    "#00C6EA", "#9B8500", "#009E73", "#FF6EB4", "#7AFF33",
    "#A837D8", "#FFD700", "#00FF99", "#FF7F50", "#8A2BE2",
    "#4B0082", "#FF1493", "#32CD32", "#8B4513"
]

In [4]:

adata = sc.read_visium(path='../data/Adult_Mouse_Brain/Section1', count_file='V1_Adult_Mouse_Brain_Coronal_Section_1_filtered_feature_bc_matrix.h5')
adata.var_names_make_unique()

In [5]:

adata

Out[5]:

AnnData object with n_obs × n_vars = 2903 × 32285
    obs: 'in_tissue', 'array_row', 'array_col'
    var: 'gene_ids', 'feature_types', 'genome'
    uns: 'spatial'
    obsm: 'spatial'

In [6]:

from sklearn.decomposition import PCA
#Normalization
sc.pp.highly_variable_genes(adata, flavor="seurat_v3", n_top_genes=2000)
adata = adata[:, adata.var['highly_variable']]
sc.pp.scale(adata)

adata_X = PCA(n_components=200, random_state=42).fit_transform(adata.X)
adata.obsm['X_pca'] = adata_X

Constructing the spatial network¶

In [7]:

# construct graph
graph_dict = HiSTaR.graph_construction(adata, 12)

Running HiSTaR¶

In [8]:

device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')
histar_net = HiSTaR.histar(adata.obsm['X_pca'], graph_dict, device=device, gcn_hidden2=12, lambda_sim=0.15)
histar_net.train()
histar_feat, _, _, _ = histar_net.process()
adata.obsm['HiSTaR'] = histar_feat

Clusting¶

In [9]:

#HiSTaR.configure_r_environment()   # If you encounter problems loading R packages, you can manually configure your path in this function.

In [10]:

HiSTaR.mclust_R(adata, n_clusters=15, use_rep='HiSTaR', key_added='HiSTaR')

R[write to console]:                    __           __ 
   ____ ___  _____/ /_  _______/ /_
  / __ `__ \/ ___/ / / / / ___/ __/
 / / / / / / /__/ / /_/ (__  ) /_  
/_/ /_/ /_/\___/_/\__,_/____/\__/   version 6.1.1
Type 'citation("mclust")' for citing this R package in publications.

fitting ...
  |======================================================================| 100%

Out[10]:

AnnData object with n_obs × n_vars = 2903 × 2000
    obs: 'in_tissue', 'array_row', 'array_col', 'HiSTaR'
    var: 'gene_ids', 'feature_types', 'genome', 'highly_variable', 'highly_variable_rank', 'means', 'variances', 'variances_norm', 'mean', 'std'
    uns: 'spatial', 'hvg'
    obsm: 'spatial', 'X_pca', 'HiSTaR'

Visualization¶

In [11]:

plt.rcParams.update({
    "figure.figsize": (4, 4),
    "figure.facecolor": "white",  
    "axes.facecolor": "white"    
})

# 生成无背景的嵌入图
sc.pl.spatial(
    adata,
    basis="spatial",
    img_key=None, 
    color="HiSTaR",
    size=1.5,
    palette=mclust_palette,
    title="HiSTaR Clustering (Spatial)",

    frameon=True,            
    legend_loc=None,         
    show=False,              
)

ax = plt.gca()
ax.set_facecolor("white")
plt.tight_layout()

In [12]:

sc.pp.neighbors(adata, use_rep='HiSTaR')
sc.tl.umap(adata)
plt.rcParams["figure.figsize"] = (4, 3)
sc.pl.umap(
    adata,
    color='HiSTaR',
    title='HiSTaR',
    show=False,
    s=16,
    palette=mclust_palette,
    frameon=False
)

Out[12]:

<Axes: title={'left': 'HiSTaR'}, xlabel='UMAP1', ylabel='UMAP2'>