Tutorial 4: Mouse Olfactory Bulb (Stereo-seq)
Import necessary packages
from sklearn import metrics
import torch
import copy
import os
import random
import numpy as np
from semanticst.loading_batches import PrepareDataloader
from semanticst.loading_batches import Dataloader
import scanpy as sc
import matplotlib.pyplot as plt
import pandas as pd
from pathlib import Path
import torch.utils.data as data
from semanticst.main import Config
/home/roxana/anaconda3/envs/semanticst3/lib/python3.9/site-packages/torch/__config__.py:10: UserWarning: CUDA initialization: The NVIDIA driver on your system is too old (found version 11040). Please update your GPU driver by downloading and installing a new version from the URL: http://www.nvidia.com/Download/index.aspx Alternatively, go to: https://pytorch.org to install a PyTorch version that has been compiled with your version of the CUDA driver. (Triggered internally at ../c10/cuda/CUDAFunctions.cpp:108.)
return torch._C._show_config()
Read data and import device
device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')
print(f"You are using *{device}*")
BASE_PATH = Path('/home/roxana/Projects/Data/5.Mouse_Olfactory/filtered_feature_bc_matrix.h5ad')
spot_paths= Path(f'{BASE_PATH}')
You are using *cpu*
dataset="mouse olfactory bulb"
adata = sc.read_h5ad(spot_paths)
adata.var_names_make_unique()
sc.pp.filter_cells(adata, min_genes=20)
sc.pp.filter_genes(adata, min_cells=50)
print(adata)
AnnData object with n_obs × n_vars = 19047 × 14376
obs: 'n_genes_by_counts', 'log1p_n_genes_by_counts', 'total_counts', 'log1p_total_counts', 'pct_counts_in_top_50_genes', 'pct_counts_in_top_100_genes', 'pct_counts_in_top_200_genes', 'pct_counts_in_top_500_genes', 'n_genes'
var: 'n_cells_by_counts', 'mean_counts', 'log1p_mean_counts', 'pct_dropout_by_counts', 'total_counts', 'log1p_total_counts', 'n_cells'
obsm: 'spatial'
plt.rcParams["figure.figsize"] = (5,4)
sc.pl.embedding(adata, basis="spatial", color="n_genes_by_counts", show=False)
plt.title("")
plt.axis('off')
(6005.190789473684, 12428.6600877193, 9986.774763741741, 15062.302776957435)
Train the model
Available data types: ‘’Xenium’, ‘Visium’, ‘Stereo’, Slide.
It is essential to specify the type of ST data, as different ST technologies require distinct preprocessing steps. Additionally, you have the option to select between mini-batch training for large datasets and full dataset training for smaller ones, ensuring efficient data processing and model performance.
dtype = "Stereo"
config=Config(device=device,dtype=dtype, use_mini_batch=False)
from semanticst.SemanticST_main import Semantic as Trainer
config_used = copy.copy(config)
model = Trainer(adata,config)
adata=model.train() # Train the model
🚀 Welcome to SemanticST! 🚀
📢 Recommendation: If your dataset contains more than 40000 spots or cells, we suggest using **mini-batch training** for efficiency.
✅ Using Full Dataset Training (No Mini-Batching). 🔥
/home/roxana/anaconda3/envs/semanticst3/lib/python3.9/site-packages/scanpy/preprocessing/_highly_variable_genes.py:75: UserWarning: `flavor='seurat_v3'` expects raw count data, but non-integers were found.
warnings.warn(
Begin to train ST data...
building sparse Matrix
127939
Learning Semantic graphs: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 250/250 [00:38<00:00, 6.53epoch/s]
Semantic Graph Learning Completed
Feature Learning Epochs: 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1000/1000 [35:23<00:00, 2.12s/it]
Clustering
- Leiden
- Louvain
from sklearn.decomposition import PCA
pca = PCA(n_components=20, random_state=1)
embedding = pca.fit_transform(adata.obsm['emb_decoder'].copy())
adata.obsm['emb_pca'] = embedding
sc.pp.neighbors(adata, use_rep='emb_pca')
sc.tl.umap(adata)
sc.tl.louvain(adata, random_state=0,resolution=1)
sc.tl.leiden(adata, random_state=0, resolution=1)
adata.uns['louvain_colors'] = ["#E31A1C","#268785","#DAB370","#C798EE","#59BE86","#006400","#8470FF",
"#CD69C9"]
plt.rcParams["figure.figsize"] = (3, 3)
fig, ax = plt.subplots()
sc.pl.embedding(adata, basis="spatial", color="louvain",s=6,palette=adata.uns['louvain_colors'], ax=ax,show=False, title='SemanticST')
plt.axis('off')
fig.savefig("sEMANTIC_seqfish_lov1_mouse_brain.png", dpi=600,bbox_inches='tight')
plt.show()
/home/tower2/anaconda3/envs/new_stellar/lib/python3.9/site-packages/scanpy/plotting/_tools/scatterplots.py:1251: FutureWarning: The default value of 'ignore' for the `na_action` parameter in pandas.Categorical.map is deprecated and will be changed to 'None' in a future version. Please set na_action to the desired value to avoid seeing this warning
color_vector = pd.Categorical(values.map(color_map))
/home/tower2/anaconda3/envs/new_stellar/lib/python3.9/site-packages/scanpy/plotting/_tools/scatterplots.py:394: UserWarning: No data for colormapping provided via 'c'. Parameters 'cmap' will be ignored
cax = scatter(
Layers’s marker genes
import matplotlib.pyplot as plt
from scipy.sparse import issparse
# Define the list of marker genes you want to plot
marker_genes = ['Mbp', 'Nrgn', 'Pcp4', 'Gabra1', 'Slc6a11', 'Cck', 'Apod']
adata.obsm['spatial'][:, 1] = -1*adata.obsm['spatial'][:, 1]
# Create a figure with subplots - one for each marker gene
fig, axes = plt.subplots(1, len(marker_genes), figsize=(20, 3)) # Adjust the figsize as needed
for i, gene in enumerate(marker_genes):
# Check if the gene is in the dataset
if gene in adata.var_names:
# Extract the expression values for the gene
gene_expression = adata[:, gene].X
# If the data is stored as a sparse matrix, convert it to a dense array
if issparse(gene_expression):
gene_expression = gene_expression.toarray().flatten()
else:
gene_expression = gene_expression.flatten()
# Invert the y-axis of the spatial coordinates for visualization (if needed)
spatial_coords = np.copy(adata.obsm['spatial'])
spatial_coords[:, 1] = -1 * spatial_coords[:, 1]
# Create the scatter plot using matplotlib for the current gene
ax = axes[i]
scatter = ax.scatter(spatial_coords[:, 0], spatial_coords[:, 1],
c=gene_expression, cmap='viridis', s=1)
ax.set_title(gene)
ax.axis('off')
# Add a color bar to each subplot if desired
# fig.colorbar(scatter, ax=ax, label='Expression Level')
else:
axes[i].text(0.5, 0.5, "Gene '{}' not found".format(gene),
transform=axes[i].transAxes,
horizontalalignment='center', verticalalignment='center')
axes[i].axis('off')
plt.colorbar(scatter, label='Expression')
# Adjust the layout so that subplots fit into the figure area
plt.tight_layout()
#plt.savefig("MARKERS_stereoseq.png", dpi=600,bbox_inches='tight')
# Show the plot
plt.show()
Clustering with mclust
n_cluster=7
tool = 'mclust' # mclust, leiden, and louvain
# clustering
from semanticst.utils import clustering
clustering(adata,seed=1, n_clusters=n_cluster, method=tool,key='emb_decoder')
fitting ...
|======================================================================| 100%
plt.rcParams["figure.figsize"] = (3,3)
plot_color=["#E31A1C","#268785","#DAB370","#C798EE","#59BE86","#006400","#8470FF",
"#CD69C9"]
sc.pl.embedding(adata, basis="spatial", color="domain",s=6,palette=plot_color, show=False, title='SemanticST')
plt.axis('off')
plt.savefig("sEMANTIC_seqfish_mclust_mouse_brain.png", dpi=600,bbox_inches='tight')
plt.show()
/home/tower2/anaconda3/envs/new_stellar/lib/python3.9/site-packages/scanpy/plotting/_tools/scatterplots.py:1251: FutureWarning: The default value of 'ignore' for the `na_action` parameter in pandas.Categorical.map is deprecated and will be changed to 'None' in a future version. Please set na_action to the desired value to avoid seeing this warning
color_vector = pd.Categorical(values.map(color_map))
/home/tower2/anaconda3/envs/new_stellar/lib/python3.9/site-packages/scanpy/plotting/_tools/scatterplots.py:394: UserWarning: No data for colormapping provided via 'c'. Parameters 'cmap' will be ignored
cax = scatter(
import matplotlib.pyplot as plt
import scanpy as sc
# Assuming adata is already defined and preprocessed
adata.obsm['spatial'][:, 1] = -1 * adata.obsm['spatial'][:, 1]
plt.rcParams["figure.figsize"] = (3,3)
# Define your custom color palette
plot_color=["#E31A1C","#268785","#DAB370","#C798EE","#59BE86","#006400","#8470FF",
"#CD69C9"]
# Ensure unique_labels order matches the intended color assignment
unique_labels = sorted(adata.obs['domain'].unique())
# Explicitly map labels to colors
label_to_color = {label: plot_color[i % len(plot_color)] for i, label in enumerate(unique_labels)}
# Apply the color mapping directly for the combined plot
sc.pl.embedding(adata, basis="spatial",
color="domain",
s=5,
palette=label_to_color, # Use the label_to_color mapping here directly
title="",
show=False)
# Plot each cluster with consistent coloring
for label in unique_labels:
fig, ax = plt.subplots()
# Plot all cells in light gray
ax.scatter(adata.obsm['spatial'][:, 0], adata.obsm['spatial'][:, 1],
c='lightgray', s=0.5)
# Overlay the cells of the current label in their respective color
label_indices = adata.obs['domain'] == label
ax.scatter(adata.obsm['spatial'][label_indices, 0],
adata.obsm['spatial'][label_indices, 1],
c=[label_to_color[label]], s=0.5) # Ensure color is in a list for consistency
ax.axis('off')
plt.savefig(f"cluster_{label}_plot.png", dpi=300, bbox_inches='tight')
plt.show()
/home/tower2/anaconda3/envs/new_stellar/lib/python3.9/site-packages/scanpy/plotting/_tools/scatterplots.py:1251: FutureWarning: The default value of 'ignore' for the `na_action` parameter in pandas.Categorical.map is deprecated and will be changed to 'None' in a future version. Please set na_action to the desired value to avoid seeing this warning
color_vector = pd.Categorical(values.map(color_map))
/home/tower2/anaconda3/envs/new_stellar/lib/python3.9/site-packages/scanpy/plotting/_tools/scatterplots.py:394: UserWarning: No data for colormapping provided via 'c'. Parameters 'cmap' will be ignored
cax = scatter(
