-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathtest_imputation.py
148 lines (114 loc) · 4.62 KB
/
test_imputation.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
import warnings
warnings.filterwarnings("ignore")
import os
import torch
import sys
from options import Options
from dataset import ImputationDataset
from tqdm import tqdm
import numpy as np
import scanpy as sc
import pandas as pd
from utils import *
from SpaIM import ImputeModule
# 预测 SC 中存在但 ST 中确实的 Genes
opt = Options().parse()
opt.kfold = 0 # 选择使用第几个fold的模型
dataset = 'Dataset1'
# 选择对应数据集
st_path = f'./dataset/{dataset}/Insitu_count.h5ad'
sc_path = f'./dataset/{dataset}/scRNA_count_cluster.h5ad'
ST_adata = sc.read(st_path)
SC_adata = sc.read(sc_path)
print(ST_adata, '\n', SC_adata)
valdataset = ImputationDataset(opt, istrain='val')
gene_names, cell_names = valdataset.get_eval_names()
valdataloader = torch.utils.data.DataLoader(
valdataset,
batch_size=opt.batch_size,
shuffle=False,
num_workers=0
)
opt.sc_dim = valdataset.get_cluster_dim()
model = ImputeModule(opt)
if opt.parallel:
model = torch.nn.DataParallel(model).cuda().module
else:
model = model.to(torch.device('cuda:%d'%(opt.gpu)))
model_path = f'./SpaIM_results/{dataset}/last_%d.pth'%(opt.kfold)
print('model_path:',model_path)
model.load(model_path)
with torch.no_grad():
eval_result = None
input_result = None
for i, (seq, st_style, _, _) in enumerate(valdataloader):
inputs = {
'scx': seq,
'st_style': st_style
}
# print('SCX:', inputs['scx'].shape, len(st_style))
model.set_input(inputs, istrain=0)
out = model.inference()
impute_result = out['st_fake'].detach().cpu().numpy()
print("impute_result",impute_result.shape) # (3482, 67798) torch.Size([3482, 67798])
eval_result = impute_result if eval_result is None else np.concatenate((eval_result, impute_result), axis=0)
print("eval_result", eval_result.shape) # (3482, 77890) (3482, 77890)
eval_result = eval_result.T
# print(eval_result[0][:10])
eval_result[eval_result <0] = 0
# print(gene_names.shape) # 检查 gene_names 的形状
# print(cell_names.shape) # 检查 cell_names 的形状
df1 = pd.DataFrame(eval_result, index=cell_names, columns=gene_names)
df1.to_pickle(os.path.join(opt.save_path, 'impute_sc_result_%d.pkl'%(opt.kfold)))
# # 读取数据，进行对比
# adata = sc.read("./Insitu_count.h5ad")
# SpaIM_adata1 = pd.read_pickle('./SpaIM/impute_sc_result_0.pkl')
# Tangram_adata2 = pd.read_pickle('./Tangram/impute_sc_result_0.pkl')
# StDiff_adata2 = pd.read_pickle("./StDiff/impute_sc_result_0.pkl")
# # 提取表达矩阵
# raw = adata.to_df()
# spaim = SpaIM_adata1
# stdiff = StDiff_adata2
# tangram = Tangram_adata2
# # 修改行名
# raw.index = ['cell' + str(i) for i in range(1, len(raw) + 1)]
# spaim.index = ['cell' + str(i) for i in range(1, len(spaim) + 1)]
# stdiff.index = ['cell' + str(i) for i in range(1, len(stdiff) + 1)]
# tangram.index = ['cell' + str(i) for i in range(1, len(tangram) + 1)]
# # 修改列名
# raw.columns = raw.columns.str.upper()
# spaim.columns = spaim.columns.str.upper()
# stdiff.columns = stdiff.columns.str.upper()
# tangram.columns = tangram.columns.str.upper()
# # 计算相关系数
# genes = ['SOX4', 'TYK2', 'GPX1', 'EZH2']
# for gene in genes:
# print(f"PCC between raw and spaim for gene {gene}:", np.corrcoef(raw[gene], spaim[gene])[0, 1])
# print(f"PCC between raw and stdiff for gene {gene}:", np.corrcoef(raw[gene], stdiff[gene])[0, 1])
# print(f"PCC between raw and tangram for gene {gene}:", np.corrcoef(raw[gene], tangram[gene])[0, 1])
# # print('\n')
# # 读取数据
# adata = sc.read("dataset/nano9-1/Insitu_count.h5ad")
# SpaIM_adata1 = df1
# # 提取表达矩阵
# raw = adata.to_df()
# spaim = SpaIM_adata1
# # stdiff = StDiff_adata2.to_df()
# # tangram = Tangram_adata2.to_df()
# # 修改行名
# raw.index = ['cell' + str(i) for i in range(1, len(raw) + 1)]
# spaim.index = ['cell' + str(i) for i in range(1, len(spaim) + 1)]
# # stdiff.index = ['cell' + str(i) for i in range(1, len(stdiff) + 1)]
# # tangram.index = ['cell' + str(i) for i in range(1, len(tangram) + 1)]
# # 修改列名
# raw.columns = raw.columns.str.upper()
# spaim.columns = spaim.columns.str.upper()
# # stdiff.columns = stdiff.columns.str.upper()
# # tangram.columns = tangram.columns.str.upper()
# # 计算相关系数
# genes = ['SOX4', 'TYK2', 'GPX1', 'EZH2']
# for gene in genes:
# print(f"PCC between raw and spaim for gene {gene}:", np.corrcoef(raw[gene], spaim[gene])[0, 1])
# # print(f"PCC between raw and stdiff for gene {gene}:", np.corrcoef(raw[gene], stdiff[gene])[0, 1])
# # print(f"PCC between raw and tangram for gene {gene}:", np.corrcoef(raw[gene], tangram[gene])[0, 1])
# print('\n')