Tutorial 2 plus: Integrating adjacent DLPFC slices with one slice has annotation

This tutorial demonstrates semi-SMILE’s ablility to integrate two adjacent slices (151674 and 151675). The slices are sampled from human dorsolateral prefrontal cortex (DLPFC) and the processed data can be downloaded from https://figshare.com/articles/dataset/DLPFC_slices_and_reference_scRNA-seq_data/27987548

import warnings
warnings.filterwarnings('ignore')

from stSMILE import SMILE

import scanpy as sc
import anndata as ad
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import squidpy as sq
import scipy.sparse as sp
from scipy import sparse
from scipy.sparse import csr_matrix
import math
import torch
import torch.nn as nn
import time
import torch.nn.functional as F
from itertools import chain
from scanpy import read_10x_h5
import torch.optim as optim
import sklearn
from sklearn.neighbors import kneighbors_graph
import gudhi
import networkx as nx
from torch_geometric.nn import GCNConv
import random
import os
import json 
import matplotlib.image as mpimg

Load data

section_ids = ['151674','151675']

def label_to_int(adataA, label_list, label_name):
    adata_label = np.array(adataA.obs[label_name].copy())
    for i in range(len(label_list)):
        need_index = np.where(adataA.obs[label_name]==label_list[i])[0]
        if len(need_index):
            adata_label[need_index] = i
    adataA.obs['ref'] = adata_label
    return adataA

adata_l = []
for i in range(len(section_ids)):
    adata_i = sc.read_h5ad('./dataset/DLPFC/DLPFC_'+ section_ids[i]+'_ST_final.h5ad')
    adata_i.obs_names = [x+'_'+section_ids[i] for x in adata_i.obs_names]
    adata_l.append(adata_i)

# convert label to int
label_list = ['Layer1', 'Layer2', 'Layer3', 'Layer4', 'Layer5', 'Layer6', 'WM']

for i in range(len(section_ids)):
    adata_l[i] = label_to_int(adata_l[i], label_list, 'Ground Truth')

adata0_sc = sc.read_h5ad('./dataset/DLPFC/DLPFC_sc_final.h5ad') 

adata0_sc

AnnData object with n_obs × n_vars = 19764 × 3010
    obs: 'orig.ident', 'nCount_RNA', 'nFeature_RNA', 'cell_type', 'cell_subtype', 'subject', 'condition', 'batch', 'n_genes', 'ref'
    var: 'features', 'n_cells', 'n_counts'
    uns: 'rank_genes_groups'
    obsm: 'X_pca'
    obsp: 'adj_f'

label0_list = adata0_sc.obs['cell_subtype'].tolist()
adata0_label_new = adata0_sc.obs['cell_subtype'].tolist()
for i in range(len(label0_list)):
    need_index = np.where(adata0_sc.obs['cell_subtype'] == label0_list[i])[0]    
    if len(need_index):
        for p in range(len(need_index)):
            adata0_label_new[need_index[p]] = i  

adata0_sc.obs['ref'] = pd.Series(adata0_label_new, index = adata0_sc.obs['cell_subtype'].index)
adata0_sc.obs['ref'] = adata0_sc.obs['ref'].astype(str)
adata0_sc.obs['ref'] = adata0_sc.obs['ref'].astype('category')
adata0_sc.obs['Ground Truth'] = adata0_sc.obs['cell_subtype']

adata_l.append(adata0_sc)
len(adata_l)

3

Run SMILE

tag_l = ['ST_ref','ST','single cell']
in_features = len(adata_l[0].var.index)
hidden_features = 512
out_features = 50
feature_method = 'GCNConv'
alpha = 0.001
beta = 1 
lamb = 0.01 
theta = 0.001 
gamma = 10 # reconstruct 
spatial_regularization_strength= 0.9
lr=1e-3
subepochs=100
epochs=200
max_patience=50
min_stop=20
random_seed=2024
gpu=0
regularization_acceleration=True
edge_subset_sz=1000000
add_topology = True
add_feature = False
add_image = False
add_sc = True
multiscale = True
anchor_type = None
anchors_all = False
use_rep_anchor = 'embedding'
align_method = 'MMD'
anchor_size=8000
iter_comb= None
n_clusters_l = [7]
edge_weights = [1,0.1,0.1]
class_rep = 'reconstruct'

adata_l = SMILE(adata_l, tag_l, section_ids, multiscale,  n_clusters_l, in_features, feature_method, hidden_features, out_features, iter_comb, anchors_all, use_rep_anchor, alpha, beta, lamb, theta, gamma,edge_weights, add_topology, add_feature, add_image, add_sc, spatial_regularization_strength, lr=lr, subepochs=subepochs, epochs=epochs, class_rep = class_rep)

Pretraining to extract embeddings of spots...
epoch   0: train spatial C loss: 1.9644, train F loss: 38.8634,
epoch  10: train spatial C loss: 0.4374, train F loss: 30.2184,
epoch  20: train spatial C loss: 0.2435, train F loss: 24.9033,
epoch  30: train spatial C loss: 0.1987, train F loss: 21.5896,
epoch  40: train spatial C loss: 0.1731, train F loss: 19.4320,
epoch  50: train spatial C loss: 0.1590, train F loss: 17.9552,
epoch  60: train spatial C loss: 0.1484, train F loss: 16.8943,
epoch  70: train spatial C loss: 0.1384, train F loss: 16.0920,
epoch  80: train spatial C loss: 0.1306, train F loss: 15.4624,
epoch  90: train spatial C loss: 0.1193, train F loss: 14.9472,
spatial data classification: Avg Accuracy = 95.405775%
Training classifier...
Training classifier...
epoch   0: overall loss: 9.3433,sc classifier loss: 3.5075,representation loss: 0.5836,within spatial regularization loss: 0.1017
epoch  10: overall loss: 3.1604,sc classifier loss: 2.2548,representation loss: 0.0906,within spatial regularization loss: 0.0657
epoch  20: overall loss: 2.2735,sc classifier loss: 1.7549,representation loss: 0.0518,within spatial regularization loss: 0.0760
epoch  30: overall loss: 1.8128,sc classifier loss: 1.3729,representation loss: 0.0440,within spatial regularization loss: 0.0916
epoch  40: overall loss: 1.4819,sc classifier loss: 1.0856,representation loss: 0.0396,within spatial regularization loss: 0.1009
epoch  50: overall loss: 1.2719,sc classifier loss: 0.8814,representation loss: 0.0390,within spatial regularization loss: 0.1067
epoch  60: overall loss: 1.1300,sc classifier loss: 0.7338,representation loss: 0.0396,within spatial regularization loss: 0.1048
epoch  70: overall loss: 1.0427,sc classifier loss: 0.6323,representation loss: 0.0410,within spatial regularization loss: 0.1096
epoch  80: overall loss: 0.9401,sc classifier loss: 0.5597,representation loss: 0.0380,within spatial regularization loss: 0.1060
epoch  90: overall loss: 0.8475,sc classifier loss: 0.5016,representation loss: 0.0346,within spatial regularization loss: 0.1101
epoch 100: overall loss: 0.7889,sc classifier loss: 0.4497,representation loss: 0.0339,within spatial regularization loss: 0.1111
epoch 110: overall loss: 0.7915,sc classifier loss: 0.4107,representation loss: 0.0381,within spatial regularization loss: 0.1129
epoch 120: overall loss: 0.7123,sc classifier loss: 0.3794,representation loss: 0.0333,within spatial regularization loss: 0.1124
epoch 130: overall loss: 0.6761,sc classifier loss: 0.3556,representation loss: 0.0320,within spatial regularization loss: 0.1119
epoch 140: overall loss: 0.6583,sc classifier loss: 0.3323,representation loss: 0.0326,within spatial regularization loss: 0.1123
epoch 150: overall loss: 0.7115,sc classifier loss: 0.3278,representation loss: 0.0384,within spatial regularization loss: 0.1064
epoch 160: overall loss: 0.8094,sc classifier loss: 0.3252,representation loss: 0.0484,within spatial regularization loss: 0.1026
epoch 170: overall loss: 0.6518,sc classifier loss: 0.2978,representation loss: 0.0354,within spatial regularization loss: 0.1112
epoch 180: overall loss: 0.5971,sc classifier loss: 0.2805,representation loss: 0.0317,within spatial regularization loss: 0.1119
epoch 190: overall loss: 0.5805,sc classifier loss: 0.2646,representation loss: 0.0316,within spatial regularization loss: 0.1114
single cell data classification: Avg Accuracy = 93.326253%


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



fitting ...
  |======================================================================| 100%
Identifying anchors...
Processing datasets (0, 1)
0.8442973272034443
The ratio of filtered mnn pairs: 0.8366676770738608
Aligning by anchors...
epoch 100: total loss:14.5369, train F loss: 14.5250, train C loss: 1.9445, train D loss: 0.0120
epoch 110: total loss:14.1393, train F loss: 14.1331, train C loss: 0.2266, train D loss: 0.0063
epoch 120: total loss:13.7923, train F loss: 13.7883, train C loss: 0.1629, train D loss: 0.0040
epoch 130: total loss:13.4869, train F loss: 13.4834, train C loss: 0.1286, train D loss: 0.0034
epoch 140: total loss:13.2132, train F loss: 13.2100, train C loss: 0.1162, train D loss: 0.0031
epoch 150: total loss:12.9538, train F loss: 12.9508, train C loss: 0.1033, train D loss: 0.0030
epoch 160: total loss:12.7292, train F loss: 12.7259, train C loss: 0.0968, train D loss: 0.0033
epoch 170: total loss:12.5311, train F loss: 12.5278, train C loss: 0.0962, train D loss: 0.0033
epoch 180: total loss:12.3626, train F loss: 12.3591, train C loss: 0.0824, train D loss: 0.0036
epoch 190: total loss:12.2245, train F loss: 12.2207, train C loss: 0.0805, train D loss: 0.0038
spatial data classification: Avg Accuracy = 98.294359%
Updating classifier...
Training classifier...
torch.Size([3635, 33])
torch.Size([3566, 33])
epoch   0: overall loss: 15.6703,sc classifier loss: 3.5437,representation loss: 1.2127,within spatial regularization loss: 0.0959
epoch  10: overall loss: 3.9806,sc classifier loss: 2.2373,representation loss: 0.1743,within spatial regularization loss: 0.0616
epoch  20: overall loss: 2.7211,sc classifier loss: 1.7246,representation loss: 0.0997,within spatial regularization loss: 0.0602
epoch  30: overall loss: 2.2522,sc classifier loss: 1.3839,representation loss: 0.0868,within spatial regularization loss: 0.0750
epoch  40: overall loss: 1.9173,sc classifier loss: 1.1276,representation loss: 0.0790,within spatial regularization loss: 0.0827
epoch  50: overall loss: 1.6932,sc classifier loss: 0.9299,representation loss: 0.0763,within spatial regularization loss: 0.0861
epoch  60: overall loss: 1.5083,sc classifier loss: 0.7830,representation loss: 0.0725,within spatial regularization loss: 0.0883
epoch  70: overall loss: 1.4289,sc classifier loss: 0.6747,representation loss: 0.0754,within spatial regularization loss: 0.0884
epoch  80: overall loss: 1.3784,sc classifier loss: 0.6001,representation loss: 0.0778,within spatial regularization loss: 0.0909
epoch  90: overall loss: 1.2288,sc classifier loss: 0.5419,representation loss: 0.0687,within spatial regularization loss: 0.0888
epoch 100: overall loss: 1.1528,sc classifier loss: 0.4884,representation loss: 0.0664,within spatial regularization loss: 0.0904
epoch 110: overall loss: 1.1057,sc classifier loss: 0.4501,representation loss: 0.0656,within spatial regularization loss: 0.0917
epoch 120: overall loss: 1.0087,sc classifier loss: 0.4157,representation loss: 0.0593,within spatial regularization loss: 0.0917
epoch 130: overall loss: 0.9706,sc classifier loss: 0.3865,representation loss: 0.0584,within spatial regularization loss: 0.0934
epoch 140: overall loss: 0.9384,sc classifier loss: 0.3609,representation loss: 0.0577,within spatial regularization loss: 0.0938
epoch 150: overall loss: 0.9120,sc classifier loss: 0.3381,representation loss: 0.0574,within spatial regularization loss: 0.0950
epoch 160: overall loss: 0.8995,sc classifier loss: 0.3188,representation loss: 0.0581,within spatial regularization loss: 0.0945
epoch 170: overall loss: 0.9075,sc classifier loss: 0.3011,representation loss: 0.0606,within spatial regularization loss: 0.0950
epoch 180: overall loss: 0.9742,sc classifier loss: 0.2911,representation loss: 0.0683,within spatial regularization loss: 0.0979
epoch 190: overall loss: 0.8948,sc classifier loss: 0.2773,representation loss: 0.0617,within spatial regularization loss: 0.0923
single cell data classification: Avg Accuracy = 92.880994%

adata_concat_st = ad.concat(adata_l[0:len(section_ids)], label="slice_name", keys=section_ids)

sc.tl.pca(adata_concat_st)
adata_concat_st.obsm['X_pca_old'] = adata_concat_st.obsm['X_pca'].copy()
adata_concat_st.obsm['X_pca'] = adata_concat_st.obsm['embedding'].copy()
sc.pp.neighbors(adata_concat_st)  
sc.tl.umap(adata_concat_st)

sc.tl.leiden(adata_concat_st, random_state=666, key_added="leiden", resolution=0.18)
len(list(set(adata_concat_st.obs['leiden'].tolist())))

7

Results and visualizations

from stSMILE import analysis
analysis.mclust_R(adata_concat_st, num_cluster=7, used_obsm="embedding")

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





AnnData object with n_obs × n_vars = 7201 × 3010
    obs: 'in_tissue', 'array_row', 'array_col', 'Ground Truth', 'n_genes', 'image_cluster', 'dbscan_cluster_new', 'ref', 'st_pre', 'slice_name', 'leiden', 'mclust'
    uns: 'pca', 'neighbors', 'umap', 'leiden'
    obsm: 'X_pca', 'features', 'features_summary_scale0.5_0.5', 'features_summary_scale0.5_1', 'features_summary_scale0.5_2', 'features_summary_scale1_0.5', 'features_summary_scale1_1', 'features_summary_scale1_2', 'features_summary_scale2_0.5', 'features_summary_scale2_1', 'features_summary_scale2_2', 'spatial', 'embedding', 'hidden_spatial', 'reconstruct', 'deconvolution', 'X_pca_old', 'X_umap'
    varm: 'PCs'
    obsp: 'distances', 'connectivities'

plt.rcParams["figure.figsize"] = (4, 4)
sc.pl.umap(adata_concat_st,color=["mclust",'Ground Truth',"slice_name"], wspace=0.4, save = 'DLPFC_umap_cluster_semiSMILE.pdf')  

WARNING: saving figure to file figures/umapDLPFC_umap_cluster_semiSMILE.pdf

import matplotlib.pyplot as plt
from sklearn.metrics import adjusted_rand_score as ari_score
from sklearn.metrics import normalized_mutual_info_score as nmi_score
# split to each data
Batch_list = []
for section_id in section_ids:
    Batch_list.append(adata_concat_st[adata_concat_st.obs['slice_name'] == section_id])

spot_size = 200
title_size = 12
ARI_list = []
NMI_list = []
for bb in range(len(section_ids)):
    ARI_list.append(round(ari_score(Batch_list[bb].obs['Ground Truth'], Batch_list[bb].obs['mclust']), 2))
    NMI_list.append(round(nmi_score(Batch_list[bb].obs['Ground Truth'], Batch_list[bb].obs['mclust']), 2))

fig, ax = plt.subplots(2, 1, figsize=(3.5, 7), gridspec_kw={'wspace': 0.05, 'hspace': 0.1})
_sc_0 = sc.pl.spatial(Batch_list[0], img_key=None, color=['mclust'], title=[''],
                      legend_loc=None, legend_fontsize=12, show=False, ax=ax[0], frameon=False,
                      spot_size=spot_size)
_sc_0[0].set_title("ARI=" + str(ARI_list[0])+",NMI=" + str(NMI_list[0]), size=title_size)
_sc_1 = sc.pl.spatial(Batch_list[1], img_key=None, color=['mclust'], title=[''],
                      legend_loc=None, legend_fontsize=12, show=False, ax=ax[1], frameon=False,
                      spot_size=spot_size)
_sc_1[0].set_title("ARI=" + str(ARI_list[1])+",NMI=" + str(NMI_list[1]), size=title_size)
plt.savefig("DLPFC_spatial_semiSMILE.pdf") 
plt.show()

adata_l[0].uns['deconvolution']

	AAACAAGTATCTCCCA-1_151674_151674	AAACAATCTACTAGCA-1_151674_151674	AAACACCAATAACTGC-1_151674_151674	AAACAGAGCGACTCCT-1_151674_151674	AAACAGCTTTCAGAAG-1_151674_151674	AAACAGGGTCTATATT-1_151674_151674	AAACAGTGTTCCTGGG-1_151674_151674	AAACATTTCCCGGATT-1_151674_151674	AAACCCGAACGAAATC-1_151674_151674	AAACCGGGTAGGTACC-1_151674_151674	...	TTGTGTATGCCACCAA-1_151674_151674	TTGTGTTTCCCGAAAG-1_151674_151674	TTGTTAGCAAATTCGA-1_151674_151674	TTGTTCAGTGTGCTAC-1_151674_151674	TTGTTGTGTGTCAAGA-1_151674_151674	TTGTTTCACATCCAGG-1_151674_151674	TTGTTTCATTAGTCTA-1_151674_151674	TTGTTTCCATACAACT-1_151674_151674	TTGTTTGTATTACACG-1_151674_151674	TTGTTTGTGTAAATTC-1_151674_151674
Ex_5_L5	5.331036e-03	7.324228e-06	1.937497e-09	5.294757e-03	1.807263e-02	1.880007e-04	1.581045e-06	2.861076e-03	1.820327e-04	1.195052e-03	...	6.635312e-05	4.834505e-05	4.722993e-04	1.908829e-02	5.307228e-02	3.280701e-08	1.399956e-08	1.044812e-04	7.543200e-07	8.032936e-07
Ex_3_L4_5	7.539445e-02	6.819289e-05	1.030208e-09	1.737979e-01	1.791637e-01	2.081110e-01	1.536111e-05	4.785919e-02	3.770103e-03	6.018199e-01	...	8.773857e-03	2.299776e-03	1.031976e-01	3.292956e-01	4.178228e-01	1.540545e-08	8.910790e-09	8.421935e-03	2.524276e-05	4.358436e-07
Ex_6_L4_6	1.575898e-06	2.745482e-05	6.694957e-09	1.030967e-06	1.205636e-03	2.139607e-04	1.182888e-03	9.814599e-06	3.724547e-08	1.472958e-04	...	2.117671e-03	1.107603e-02	3.724185e-03	4.251718e-03	2.000385e-03	6.538797e-08	7.465867e-08	1.571381e-02	1.732987e-03	1.025283e-06
Inhib_5	3.887602e-05	3.755447e-03	1.494762e-09	4.206381e-06	3.983635e-06	3.805378e-08	3.293550e-08	1.616721e-03	1.622213e-08	1.483451e-08	...	1.865695e-07	5.408108e-07	6.714814e-04	5.229215e-08	7.182640e-08	2.559299e-08	1.129170e-08	1.972515e-06	4.577277e-09	6.487975e-04
Inhib_4_SST	6.003727e-06	1.779631e-03	2.024276e-10	2.864599e-06	1.316401e-04	4.011348e-07	2.751985e-08	2.927481e-05	1.876359e-07	4.262075e-07	...	4.628414e-07	9.928970e-07	8.035329e-06	2.911775e-05	4.350154e-05	7.866907e-10	8.385642e-10	2.388119e-06	5.029619e-09	4.906387e-05
OPCs_2	9.411585e-12	2.803001e-08	2.958015e-05	1.459952e-12	3.119064e-07	4.982614e-07	6.164388e-05	1.319579e-10	2.043856e-12	3.260431e-09	...	3.264469e-05	2.948959e-04	1.025248e-10	2.536996e-09	7.815594e-09	1.441980e-03	2.117293e-04	4.960734e-04	8.936862e-06	6.285780e-07
Oligos_1	1.103193e-09	1.334777e-08	6.613036e-01	1.950127e-10	1.607728e-05	5.314733e-06	1.269162e-01	3.042426e-09	1.468542e-11	3.137766e-07	...	1.934497e-04	1.861525e-03	3.766803e-11	2.829274e-09	1.265834e-08	3.549326e-01	5.432719e-01	9.517118e-04	4.933485e-02	2.862671e-10
Inhib_2_VIP	1.142907e-04	2.737476e-03	6.132087e-08	4.443296e-05	2.542165e-06	7.336904e-08	1.811521e-08	8.416994e-04	1.239800e-06	1.294661e-08	...	4.978689e-07	1.027854e-06	6.562318e-04	2.064023e-06	2.481426e-06	7.098891e-07	1.599951e-07	3.750144e-06	1.140748e-09	1.657319e-02
OPCs_1	1.060887e-06	5.427985e-06	4.150887e-06	2.897693e-07	1.265667e-05	1.501308e-05	2.545834e-05	1.609986e-06	1.650403e-06	5.230331e-07	...	3.125764e-04	7.122649e-04	9.678852e-08	1.952605e-06	7.306139e-06	1.551582e-04	1.766666e-05	1.211827e-03	7.375433e-06	1.482470e-03
Ex_4_L_6	9.386843e-07	8.146332e-07	1.562857e-08	1.397563e-06	4.251486e-01	3.273296e-01	2.726563e-05	4.943303e-07	2.445179e-06	3.013038e-01	...	4.658365e-02	1.329014e-02	1.465842e-07	1.349780e-02	5.322984e-02	1.111096e-07	7.676510e-08	3.381010e-02	3.375684e-05	7.671878e-08
Inhib_3_SST	4.578171e-06	3.401444e-02	2.195845e-08	2.991445e-06	1.034968e-05	5.285029e-06	7.356957e-06	2.125660e-05	1.916202e-06	1.477993e-06	...	3.658260e-05	1.695212e-04	1.678331e-05	3.886961e-06	5.464119e-06	8.078320e-08	9.694001e-08	1.783061e-04	3.585289e-06	7.286732e-02
Micro/Macro	2.150780e-05	3.946796e-05	9.279989e-05	3.148427e-06	8.179383e-05	3.732434e-06	9.665425e-05	1.013447e-04	1.821458e-09	4.200255e-06	...	2.788698e-05	3.078261e-05	8.546996e-07	1.906584e-09	6.599026e-09	2.650844e-05	7.879714e-05	7.170058e-05	1.959703e-05	4.396588e-09
Mix_1	2.023518e-07	4.801752e-08	1.651792e-07	4.279693e-08	6.500719e-02	9.163583e-05	3.579773e-05	8.461726e-07	6.530758e-10	1.560206e-04	...	4.352113e-05	9.424517e-05	2.463661e-08	1.416680e-06	1.149178e-05	1.209067e-06	7.319786e-07	1.364390e-04	2.047870e-05	4.668186e-10
Oligos_2	1.319824e-08	1.834519e-07	5.481540e-02	2.027216e-09	1.892705e-05	2.267293e-06	7.492923e-04	5.514750e-08	1.110849e-10	1.142498e-07	...	4.347417e-05	2.343807e-04	1.066354e-09	1.877399e-08	8.906188e-08	1.582011e-01	7.045989e-02	2.358791e-04	1.008167e-04	6.806261e-08
Inhib_1	9.456518e-06	5.382027e-06	6.668554e-08	4.203802e-06	2.071941e-05	2.205932e-08	1.627803e-08	1.251685e-05	8.950122e-09	5.379627e-08	...	7.012682e-08	1.763423e-07	2.499046e-06	5.011632e-04	7.377816e-04	5.148976e-08	5.739727e-08	3.959746e-07	9.493011e-10	3.387155e-07
Mix_5	3.775398e-02	2.983437e-04	7.938143e-06	2.217800e-02	1.719568e-03	9.524930e-05	4.407734e-05	3.544265e-02	2.640762e-04	7.582183e-05	...	2.161817e-04	4.671341e-04	2.099425e-03	1.682449e-02	3.010873e-02	5.564242e-05	2.004363e-05	1.096651e-03	1.725855e-05	5.727104e-04
Oligos_3	1.339343e-06	2.749919e-06	6.864335e-03	3.779722e-07	1.681903e-04	3.020849e-05	8.184273e-03	1.727511e-06	3.279376e-07	2.599987e-06	...	5.616409e-04	1.803871e-03	2.128524e-07	1.337305e-05	3.229907e-05	2.281751e-01	2.331905e-02	2.335216e-03	2.593313e-03	5.908923e-06
Ex_7_L4_6	1.265088e-09	1.448083e-08	1.310187e-08	1.278454e-09	2.435715e-03	2.486624e-04	4.005212e-04	1.146575e-09	2.247563e-09	2.020953e-05	...	7.581071e-04	7.062975e-03	1.216199e-09	1.270670e-02	5.825293e-02	5.130069e-07	2.386969e-07	7.764046e-03	3.676155e-04	2.093845e-08
Ex_1_L5_6	6.708034e-14	7.432730e-12	1.169388e-04	1.438007e-14	4.978217e-02	4.008515e-05	8.784932e-05	2.891945e-13	6.747332e-16	1.772819e-06	...	1.195745e-04	8.653622e-04	1.086763e-12	2.970352e-03	2.290949e-02	8.726579e-03	6.195587e-04	8.044475e-03	1.219586e-05	2.063101e-13
Mix_4	2.181848e-05	1.083354e-06	1.585618e-06	7.154743e-06	1.327700e-03	1.434868e-05	8.078919e-06	1.508239e-05	6.845541e-08	5.916182e-05	...	4.211684e-05	8.542206e-05	1.025069e-07	1.186687e-05	8.130258e-05	7.573494e-06	3.135555e-06	3.640089e-04	1.928896e-06	9.510257e-08
Inhib_6_SST	1.554437e-03	2.165699e-07	5.253943e-08	1.110127e-03	1.013113e-05	9.396200e-09	1.047670e-08	6.310049e-04	2.091033e-05	8.094664e-09	...	2.852434e-08	1.117575e-07	1.179586e-04	1.979603e-02	2.063605e-02	1.167061e-06	1.440366e-07	2.645498e-07	5.642926e-10	2.383495e-05
Ex_2_L5	1.250792e-01	1.487714e-01	2.009409e-09	1.187723e-01	2.554095e-02	7.842440e-05	2.048043e-06	1.362125e-01	1.671452e-01	2.363377e-04	...	1.753663e-05	1.818949e-05	1.038991e-03	1.365497e-04	6.925314e-04	2.095807e-08	1.111339e-08	1.985297e-05	2.102526e-06	7.547128e-02
Inhib_8_PVALB	3.137120e-04	5.000665e-01	1.009287e-08	1.461515e-04	6.011871e-04	9.916092e-06	2.214579e-05	3.225113e-03	3.473086e-05	4.203666e-06	...	1.729646e-05	4.894003e-05	9.109486e-04	6.600137e-06	1.380142e-05	6.112612e-08	6.199973e-08	4.663732e-05	1.755405e-05	1.606022e-01
Ex_10_L2_4	7.426242e-01	2.854269e-01	2.601273e-07	6.744429e-01	3.196594e-02	1.722727e-01	1.039244e-03	6.791140e-01	8.284303e-01	2.536228e-02	...	4.609047e-01	1.767566e-01	5.989573e-02	4.285125e-04	1.017905e-03	2.726523e-06	2.370724e-06	1.177525e-01	2.840737e-03	6.517859e-01
Astros_2	4.407651e-05	8.192485e-05	2.790105e-04	1.191013e-05	2.897381e-05	8.134085e-07	5.111986e-06	6.196243e-05	2.585014e-06	2.009086e-07	...	8.109741e-06	1.246517e-05	2.156572e-07	5.336699e-07	2.824991e-06	1.574462e-02	6.534528e-04	4.654202e-05	3.549511e-07	2.068751e-03
Endo	1.846407e-07	1.172838e-04	7.158233e-02	4.867097e-08	3.630085e-03	2.035400e-02	5.225666e-01	7.006853e-07	3.583206e-09	1.423539e-03	...	1.277818e-01	2.940560e-01	6.758000e-09	7.188731e-08	3.865507e-07	1.336352e-02	1.039849e-01	2.371631e-01	4.666061e-01	2.154636e-07
Astros_1	1.174598e-10	5.284930e-08	2.032328e-01	1.903077e-11	1.755139e-06	1.546321e-06	1.910865e-03	2.740917e-10	6.233921e-12	5.801396e-08	...	4.928529e-05	2.360478e-04	1.280484e-13	2.185848e-12	2.892360e-11	1.807888e-01	2.496789e-01	1.675681e-04	1.973617e-04	9.244927e-09
Inhib_7_PVALB	3.107781e-03	1.277954e-04	3.368560e-10	9.313971e-04	2.075744e-02	1.928053e-05	1.606558e-07	4.054386e-02	3.437884e-07	3.727704e-05	...	4.754766e-06	5.486481e-06	6.700169e-01	4.568524e-01	2.505307e-01	3.965241e-09	2.020975e-09	2.161855e-05	5.491373e-08	6.068345e-07
Mix_2	1.004144e-05	1.653449e-06	6.025379e-07	1.488964e-06	1.154766e-01	1.376075e-04	1.566847e-05	4.695126e-05	3.664334e-09	3.158174e-04	...	8.254055e-05	1.792856e-04	1.027101e-05	9.353810e-05	4.768217e-04	3.585676e-06	1.649501e-06	4.078890e-04	6.169567e-06	1.721464e-08
Mix_3	8.550617e-03	2.264480e-02	1.084012e-08	3.214811e-03	6.843353e-03	1.213823e-05	3.341140e-06	5.134280e-02	7.748995e-05	1.632890e-05	...	1.170016e-05	2.899700e-05	1.571581e-01	1.226534e-01	8.671648e-02	1.105583e-07	4.949691e-08	3.657339e-05	1.942507e-06	1.783023e-02
Ex_8_L5_6	2.243421e-09	6.304414e-08	4.355186e-04	1.454770e-09	4.846143e-02	2.702842e-01	3.363685e-01	3.047409e-09	9.672512e-10	6.697954e-02	...	3.508029e-01	4.874664e-01	7.399700e-11	1.185523e-06	1.446570e-05	5.523010e-03	2.653915e-03	5.624420e-01	4.760079e-01	3.900745e-09
Ex_9_L5_6	1.461552e-05	1.779963e-05	9.042719e-10	2.614222e-05	2.345920e-03	4.252439e-04	1.545127e-06	5.732153e-06	6.430119e-05	8.355253e-04	...	2.017272e-04	1.503065e-04	8.468319e-07	8.313114e-04	1.579871e-03	1.049121e-08	6.970614e-09	2.146831e-04	9.236263e-07	1.210168e-05
Astros_3	9.727123e-11	1.935676e-07	1.232625e-03	2.886797e-11	6.258192e-06	8.700958e-06	2.204419e-04	1.306050e-10	2.727591e-10	1.324405e-07	...	1.870019e-04	6.419263e-04	1.202216e-13	1.377305e-10	1.800484e-09	3.284720e-02	5.021272e-03	7.355794e-04	3.811958e-05	1.928544e-06

33 rows × 3635 columns

# write out the deconvolution result
for i in range(len(adata_l)-1):
    adata_i = adata_l[i].copy()
    dev_i = adata_i.uns['deconvolution']
    dev_i.to_csv('DLPFC_semiSMILE_dev_'+ section_ids[i]+'.csv', sep='\t')