Imputation and Integration for Mouse Brain data
loading package
[15]:
import scanpy as sc
import torch as th
import scanpy as sc
import pandas as pd
import torch.nn as nn
from SpaMIE.create_graph import Sagegraph
from SpaMIE.spamie_main import Sagewrapper
device = th.device('cuda:0' if th.cuda.is_available() else 'cpu')
file_fold = '/data/xiangdw/data/data/'
from matplotlib import rcParams
config = {
"font.family":'Times New Roman',
"font.size":12,
"axes.unicode_minus": False
}
rcParams.update(config)
Imputation
[ ]:
path = '/data/xiangdw/data/data/MouseBrain_new/'
a = []
layers_nums = 3
for i in range(1):
seeds = i+1
adata_omics2 = sc.read_h5ad(path + 'MouseBrain_RNA_concat.h5ad')
adata_omics1 = sc.read_h5ad(path + 'MouseBrain_H3K27me3_concat.h5ad')
#New spatial coord
tissue_positions = pd.read_csv('/data/xiangdw/mousebrain/gtt_output/new_coord_modify.csv', header=None)
cell_ids = tissue_positions.iloc[:, 0]
x_coords = tissue_positions.iloc[:, 1]
y_coords = tissue_positions.iloc[:, 2]
x_coords_dict = dict(zip(cell_ids, x_coords))
y_coords_dict = dict(zip(cell_ids, y_coords))
adata_omics1.obs_names = adata_omics1.obs_names.str.replace('mousebrain#', '')
adata_omics2.obs_names = adata_omics2.obs_names.str.replace('mousebrain#', '')
adata_omics1.obs['x_coord'] = adata_omics1.obs.index.map(x_coords_dict).astype('float')
adata_omics1.obs['y_coord'] = adata_omics1.obs.index.map(y_coords_dict).astype('float')
adata_omics1.obsm['spatial'] = adata_omics1.obs[['x_coord', 'y_coord']].values
adata_omics2.obs['x_coord'] = adata_omics2.obs['row'].values
adata_omics2.obs['y_coord'] = adata_omics2.obs['col'].values
adata_omics2.obsm['spatial'] = adata_omics2.obs[['x_coord', 'y_coord']].values
modalities = [adata_omics1, adata_omics2]
g_spatial_omics1, g_feature_omics1, adata_omics1, adata_omics2 = Sagegraph(modalities, device, datatype='Spatial_Epigenome_RNA', batch=True)
output_dir = '/data/xiangdw/data/pred result/'
weight = [0,1,2]
pred_name = 'MouseBrain_SpaMIE_RNA_'+str(layers_nums)+'_pred.csv'
true_name = 'MouseBrain_SpaMIE_RNA_'+str(layers_nums)+'_truth.csv'
in_feat = adata_omics1.obsm['feat'].shape[1]
out_feat = adata_omics2.X.shape[1]
train_size = adata_omics1[adata_omics1.obs['batch']=='1'].shape[0]
model = Sagewrapper(seed=(int(seeds)), device=device, in_feat=in_feat, n_hidden=256, out_feat=out_feat, task='prediction', datatype='Spatial_Epigenome_RNA',
layers_nums=int(layers_nums), weight=weight, epoch=600, res_type='res_add', activation=nn.LeakyReLU
, sagetype='mean', lr=2e-4, lr2 = 0.002)
adata_omics1_pred, adata_omics2_pred, test_idx, train_idx,wt,alph = model.fit(adata_omics1, adata_omics2, g_spatial_omics1, g_feature_omics1,
output_dir=output_dir, pred_name=pred_name,
true_name=true_name, train_size=train_size, weight=True, save_csv=False)
2025-10-17 15:29:31,977 - harmonypy - INFO - Computing initial centroids with sklearn.KMeans...
2025-10-17 15:30:16,370 - harmonypy - INFO - sklearn.KMeans initialization complete.
2025-10-17 15:30:16,514 - harmonypy - INFO - Iteration 1 of 10
2025-10-17 15:30:24,687 - harmonypy - INFO - Iteration 2 of 10
2025-10-17 15:30:33,012 - harmonypy - INFO - Iteration 3 of 10
2025-10-17 15:30:40,001 - harmonypy - INFO - Iteration 4 of 10
2025-10-17 15:30:45,364 - harmonypy - INFO - Iteration 5 of 10
2025-10-17 15:30:50,683 - harmonypy - INFO - Iteration 6 of 10
2025-10-17 15:30:56,000 - harmonypy - INFO - Iteration 7 of 10
2025-10-17 15:31:01,288 - harmonypy - INFO - Iteration 8 of 10
2025-10-17 15:31:08,336 - harmonypy - INFO - Iteration 9 of 10
2025-10-17 15:31:12,838 - harmonypy - INFO - Iteration 10 of 10
2025-10-17 15:31:17,571 - harmonypy - INFO - Stopped before convergence
/data/xiangdw/MODEL/SpaMIE/spamie_net.py:98: UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument.
self.alpha = F.softmax(torch.squeeze(self.vu) + 1e-6)
Integration
[ ]:
import os
import dgl
import pandas as pd
import sys
import scanpy as sc
import importlib
import torch as th
import torch.nn as nn
from sklearn.utils import shuffle
from model_integration import *
from matplotlib import rcParams
config = {
"font.family":'Times New Roman',
"font.size":20,
"axes.unicode_minus": False
}
rcParams.update(config)
[4]:
import torch.nn.functional as F
from SpaMIE.create_graph import Sagegraph
from SpaMIE.spamie_main import Sagewrapper
import numpy as np
from model_integration import set_seed
device = th.device('cuda:1' if th.cuda.is_available() else 'cpu')
for i in range(1):
seeds = str(i+1)
path = '/data/xiangdw/data/data/MouseBrain_new/'
adata_omics1 = sc.read_h5ad(path + 'MouseBrain_H3K27me3_concat.h5ad')
adata_omics2 = sc.read_h5ad(path + 'MouseBrain_RNA_pred_concat.h5ad')
set_seed(2024)
#New spatial coord
tissue_positions = pd.read_csv('/data/xiangdw/mousebrain/gtt_output/new_coord_modify.csv', header=None)
cell_ids = tissue_positions.iloc[:, 0]
x_coords = tissue_positions.iloc[:, 1]
y_coords = tissue_positions.iloc[:, 2]
x_coords_dict = dict(zip(cell_ids, x_coords))
y_coords_dict = dict(zip(cell_ids, y_coords))
adata_omics1.obs_names = adata_omics1.obs_names.str.replace('mousebrain#', '')
adata_omics2.obs_names = adata_omics2.obs_names.str.replace('mousebrain#', '')
adata_omics1.obs['x_coord'] = adata_omics1.obs.index.map(x_coords_dict).astype('float')
adata_omics1.obs['y_coord'] = adata_omics1.obs.index.map(y_coords_dict).astype('float')
adata_omics1.obsm['spatial'] = adata_omics1.obs[['x_coord', 'y_coord']].values
adata_omics2.obs['x_coord'] = adata_omics2.obs['row'].values
adata_omics2.obs['y_coord'] = adata_omics2.obs['col'].values
adata_omics2.obsm['spatial'] = adata_omics2.obs[['x_coord', 'y_coord']].values
modalities = [adata_omics1, adata_omics2]
g_spatial_omics1, g_feature_omics1, g_spatial_omics2, g_feature_omics2, adata_omics1, adata_omics2 = Sagegraph(modalities, device, task="Integration", datatype="Spatial_Epigenome_RNA",batch=True)
in_feat = adata_omics1.obsm['feat'].shape[1]
out_feat = adata_omics2.X.shape[1]
weight = [1,1,1]
model = Sagewrapper(seed=(int(seeds)), device=device, in_feat=in_feat, n_hidden=256, out_feat=out_feat, task='integration', datatype='Spatial_Epigenome_RNA',
layers_nums=int(3), weight=weight, epoch=600, res_type='res_add', activation=nn.LeakyReLU
, sagetype='mean', lr=2e-4, lr2 = 0.002)
output = model.fit(adata_omics1, adata_omics2, g_spatial_omics1, g_feature_omics1, g_spatial_omics2, g_feature_omics2, weight_factors=[2,1,1,1])
adata_omics2.obsm['SpaMIE'] = output[0]
2025-10-17 15:32:22,090 - harmonypy - INFO - Computing initial centroids with sklearn.KMeans...
2025-10-17 15:32:54,144 - harmonypy - INFO - sklearn.KMeans initialization complete.
2025-10-17 15:32:54,200 - harmonypy - INFO - Iteration 1 of 10
2025-10-17 15:33:02,155 - harmonypy - INFO - Iteration 2 of 10
2025-10-17 15:33:09,790 - harmonypy - INFO - Iteration 3 of 10
2025-10-17 15:33:16,496 - harmonypy - INFO - Iteration 4 of 10
2025-10-17 15:33:21,819 - harmonypy - INFO - Iteration 5 of 10
2025-10-17 15:33:27,183 - harmonypy - INFO - Iteration 6 of 10
2025-10-17 15:33:32,600 - harmonypy - INFO - Iteration 7 of 10
2025-10-17 15:33:37,952 - harmonypy - INFO - Iteration 8 of 10
2025-10-17 15:33:45,007 - harmonypy - INFO - Iteration 9 of 10
2025-10-17 15:33:50,464 - harmonypy - INFO - Iteration 10 of 10
2025-10-17 15:33:55,879 - harmonypy - INFO - Stopped before convergence
2025-10-17 15:34:14,333 - harmonypy - INFO - Computing initial centroids with sklearn.KMeans...
2025-10-17 15:34:48,374 - harmonypy - INFO - sklearn.KMeans initialization complete.
2025-10-17 15:34:48,424 - harmonypy - INFO - Iteration 1 of 10
2025-10-17 15:34:56,293 - harmonypy - INFO - Iteration 2 of 10
2025-10-17 15:35:04,031 - harmonypy - INFO - Converged after 2 iterations
/data/xiangdw/MODEL/SpaMIE/spamie_net.py:98: UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument.
self.alpha = F.softmax(torch.squeeze(self.vu) + 1e-6)
/data/xiangdw/MODEL/SpaMIE/spamie_net.py:66: UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument.
self.alpha = F.softmax(torch.squeeze(self.vu) + 1e-6)
Visualisation
[14]:
from SpatialGlue.utils import clustering
fig, axs = plt.subplots(1, 2, figsize=(14, 7), dpi=300)
plt.rcParams['font.size'] = 20
tool = 'louvain'
path = '/data/xiangdw/data/data/MouseBrain_new/'
adata = sc.read_h5ad(path + "MouseBrain_concat_methods.h5ad")
adata.obsm['spatial'] = adata.obs[['row', 'col']].values
clustering(adata_omics2, key='SpaMIE', add_key='SpaMIE',start=0.001, end=1.2, increment=0.1, n_clusters=8, method=tool, use_pca=False)
adata.obsm['SpaMIE'] = adata_omics2.obs['SpaMIE'].copy()
adata1 = adata[adata.obs['batch'] == '1'].copy()
adata2 = adata[adata.obs['batch'] == '2'].copy()
sc.pl.embedding(adata1, basis='spatial', color=['SpaMIE_pred_1'], ax=axs[0], title='SpaMIE', s=100, show=False)
sc.pl.embedding(adata2, basis='spatial', color=['SpaMIE_pred_1'], ax=axs[1], title='SpaMIE', s=300, show=False)
Searching resolution...
resolution=1.101, cluster number=9
resolution=1.001, cluster number=8
/data/xiangdw/conda_env/GNNS/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(
/data/xiangdw/conda_env/GNNS/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(
[14]:
<Axes: title={'center': 'SpaMIE'}, xlabel='spatial1', ylabel='spatial2'>
