-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathprepare_data_Celegans.py
192 lines (174 loc) · 7.98 KB
/
prepare_data_Celegans.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
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
# prepare_data_Celegans.py
#
# 1. This is a data pre-processing file for generating Celegans dataset in PyTorch format.
# 2. The dataset was not divided in its given format. Therefore, it was processed into
# five subsets, each of which acted as test set once, i.e., 5-fold.
#
# Author: Shugang Zhang
# Created: Friday, Aug 6th, 2021
# Last update: Friday, Aug 6th, 2021
import pandas as pd
import numpy as np
import os
import json,pickle
from collections import OrderedDict
from rdkit import Chem
from rdkit.Chem import MolFromSmiles
import networkx as nx
from utils import *
def atom_features(atom):
return np.array(one_of_k_encoding_unk(atom.GetSymbol(),['C', 'N', 'O', 'S', 'F', 'Si', 'P', 'Cl', 'Br', 'Mg', 'Na','Ca', 'Fe', 'As', 'Al', 'I', 'B', 'V', 'K', 'Tl', 'Yb','Sb', 'Sn', 'Ag', 'Pd', 'Co', 'Se', 'Ti', 'Zn', 'H','Li', 'Ge', 'Cu', 'Au', 'Ni', 'Cd', 'In', 'Mn', 'Zr','Cr', 'Pt', 'Hg', 'Pb', 'Unknown']) +
one_of_k_encoding(atom.GetDegree(), [0, 1, 2, 3, 4, 5, 6,7,8,9,10]) +
one_of_k_encoding_unk(atom.GetTotalNumHs(), [0, 1, 2, 3, 4, 5, 6,7,8,9,10]) +
one_of_k_encoding_unk(atom.GetImplicitValence(), [0, 1, 2, 3, 4, 5, 6,7,8,9,10]) +
[atom.GetIsAromatic()])
def one_of_k_encoding(x, allowable_set):
if x not in allowable_set:
raise Exception("input {0} not in allowable set{1}:".format(x, allowable_set))
return list(map(lambda s: x == s, allowable_set))
def one_of_k_encoding_unk(x, allowable_set):
"""Maps inputs not in the allowable set to the last element."""
if x not in allowable_set:
x = allowable_set[-1]
return list(map(lambda s: x == s, allowable_set))
def smile_to_graph(smile):
mol = Chem.MolFromSmiles(smile)
if (mol is None):
print("bad smile:", smile)
else:
# 1.num of atoms
num_atoms = mol.GetNumAtoms()
# 2.features
features = []
for atom in mol.GetAtoms():
feature = atom_features(atom)
features.append( feature / sum(feature) )
# 3.edges
edges = []
for bond in mol.GetBonds():
edges.append([bond.GetBeginAtomIdx(), bond.GetEndAtomIdx()])
g = nx.Graph(edges).to_directed()
edge_index = []
for e1, e2 in g.edges:
edge_index.append([e1, e2])
return num_atoms, features, edge_index
def seq_cat(prot):
x = np.zeros(max_seq_len)
for i, ch in enumerate(prot[:max_seq_len]):
x[i] = seq_dict[ch]
return x
datasets = ['Celegans']
for dataset in datasets:
print('convert data from DeepDTA for ', dataset)
fpath = 'data/' + dataset + '/' + dataset + '.txt'
data = pd.read_table(fpath, sep=' ', header=None)
seq_voc = "ABCDEFGHIKLMNOPQRSTUVWXYZ"
seq_dict = {v:(i+1) for i,v in enumerate(seq_voc)} # encode alphabet from 1
seq_dict_len = len(seq_dict)
max_seq_len = 1000
# 2. create graph for all SMILES
print("\nCreating graph for all SMILES...")
compound_iso_smiles = []
for dataset in datasets:
df = pd.read_table('data/' + dataset + '/' + dataset + '.txt', sep=' ', header=None)
compound_iso_smiles += list(df[0]) # the first column is drug SMILES
compound_iso_smiles = set(compound_iso_smiles) # function set() can remove redundant automatically
smile_graph = {}
for smile in compound_iso_smiles:
g = smile_to_graph(smile)
smile_graph[smile] = g
print("Finished.")
# 3. convert to PyTorch data format
print("\nConvert to pytorch data format...")
for dataset in datasets:
df_raw = pd.read_table('data/' + dataset + '/' + dataset + '.txt', sep=' ', header=None)
df = df_raw.drop(df[(df[0] == '[Zn]')
| (df[0] == '[Mg+2]')
| (df[0] == '[Fe]')
| (df[0] == '[2H]')
| (df[0] == '[Na+].[Na+].[S-2]')
| (df[0] == '[43Ca]')
| (df[0] == '[75Se]')
| (df[0] == '[78Se]')
| (df[0] == '[45K]')
| (df[0] == '[56Fe]')
| (df[0] == '[121Sn]')
| (df[0] == '[125Sb]')
| (df[0] == '[15NH3]')
| (df[0] == '[17NH3]')
| (df[0] == 'O.[Ag].[Ag]')
| (df[0] == '[F-].[F-].[Ca+2]')
| (df[0] == 'O.O.O.[Co]')
| (df[0] == 'O.O.O.O.[V]')
| (df[0] == '[NH4+].F.[F-]')
| (df[0] == '[Li+].[AlH4-]')
| (df[0] == '[SiH2]')
| (df[0] == '[O-2].[O-2].[O-2].[O-2].[O-2].[Al+3].[Al+3].[Ca+2].[Ca+2]')
| (df[0] == '[Na+].[Na+].[Cl-].[Cl-].[Cl-].[Cl-].[Pt+2]')
| (df[0] == '[Br-].[Rb+]')
| (df[0] == '[125IH]')
| (df[0] == '[127IH]')
| (df[0] == '[129IH]')
| (df[0] == '[43K]')
| (df[0] == 'S.[As]')
| (df[0] == '[Cl]')
| (df[0] == '[Br-].[Rb+]')].index)
print(len(list(df_raw[0])))
print(len(list(df[0])))
portion = int(0.2 * len(df[0]))
drugs, prots, Y = list(df[0]), list(df[1]), list(df[2])
XT = [seq_cat(t) for t in prots]
drugs, prots, Y = np.asarray(drugs), np.asarray(XT), np.asarray(Y)
for fold in range(5):
# train_drugs = np.array([])
# train_prots = np.array([])
# train_Y = np.array([])
# test_drugs = np.array([])
# test_prots = np.array([])
# test_Y = np.array([])
if fold < 4:
# test set
test_drugs = drugs[fold * portion:(fold + 1) * portion]
test_prots = prots[fold * portion:(fold + 1) * portion]
test_Y = Y[fold * portion:(fold + 1) * portion]
# train set
train_drugs = np.delete(drugs, range(fold * portion, (fold + 1) * portion))
train_prots = np.delete(prots, range(fold * portion, (fold + 1) * portion), axis=0)
train_Y = np.delete( Y, range(fold * portion, (fold + 1) * portion))
# bug check
test_drugs = drugs
test_prots = prots
test_Y = Y
train_drugs = drugs
train_prots = prots
train_Y = Y
if fold == 4: # to avoid cases that the number of samples cannot be divided by 5
# test set
test_drugs = drugs[fold * portion:]
test_prots = prots[fold * portion:]
test_Y = Y[fold * portion:]
# train set
train_drugs = np.delete(drugs, range(fold * portion, len(drugs)))
train_prots = np.delete(prots, range(fold * portion, len(drugs)), axis=0)
train_Y = np.delete( Y, range(fold * portion, len(drugs)))
# print(len(train_drugs))
# print(len(train_prots))
# print(len(test_Y))
# make data PyTorch Geometric ready
# train set
data_path = 'data/processed/' + dataset + '_train_' + str(fold+1) + '.pt'
if not os.path.isfile(data_path): # if not exists
print('preparing ', dataset + '_train_' + str(fold+1) + '.pt in pytorch format!')
TestbedDataset(root='data', dataset=dataset + '_train_' + str(fold+1),
xd=train_drugs, xt=train_prots, y=train_Y, smile_graph=smile_graph)
print(data_path, 'created successfully.')
else:
print(data_path, 'are already created.')
# test set
data_path = 'data/processed/' + dataset + '_test_' + str(fold+1) + '.pt'
if not os.path.isfile(data_path): # if not exists
print('preparing ', dataset + '_test_' + str(fold+1) + '.pt in pytorch format!')
TestbedDataset(root='data', dataset=dataset + '_test_' + str(fold+1),
xd=test_drugs, xt=test_prots, y=test_Y, smile_graph=smile_graph)
else:
print(data_path, 'are already created.')