-
Notifications
You must be signed in to change notification settings - Fork 3
/
predictor.py
146 lines (120 loc) · 3.96 KB
/
predictor.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
# -*- coding: utf-8 -*-
# @Author : twd
# @FileName: predictor.py
# @Software: PyCharm
import os
from ETFC.model import *
os.environ["CUDA_VISIBLE_DEVICES"] = "0"
import numpy as np
from pathlib import Path
import argparse
import torch
def ArgsGet():
parse = argparse.ArgumentParser(description='ETFC')
parse.add_argument('-file', type=str, default='./test_data.fasta', help='fasta file')
parse.add_argument('-out_path', type=str, default='./ETFC/result', help='output path')
args = parse.parse_args()
return args
def get_data(file):
# getting file and encoding
seqs = []
names = []
seq_length = []
with open(file) as f:
for each in f:
if each == '\n':
continue
elif each[0] == '>':
names.append(each)
else:
seqs.append(each.rstrip())
# encoding
amino_acids = 'XACDEFGHIKLMNPQRSTVWY'
max_len = 50
data_e = []
delSeq = 0
for i in range(len(seqs)):
sign = True
if len(seqs[i]) > max_len or len(seqs[i]) < 5:
print(f'本方法只能识别序列长度在5-50AA的多肽,该序列将不能识别:{seqs[i]}')
del names[i-delSeq]
delSeq += 1
continue
length = len(seqs[i])
seq_length.append(length)
elemt, st = [], seqs[i]
for j in st:
if j == ',' or j == '1' or j == '0':
continue
elif j not in amino_acids:
sign = False
print(f'本方法只能识别包含天然氨基酸的多肽,该序列不能识别{seqs[i]}')
del names[i-delSeq]
delSeq += 1
break
index = amino_acids.index(j)
elemt.append(index)
if length <= max_len and sign:
elemt += [0] * (max_len - length)
data_e.append(elemt)
return np.array(data_e), names, np.array(seq_length)
def predict(test, seq_length, h5_model):
dir = './ETFC/dataset/Model/teacher/tea_model.pth'
print('predicting...')
# 1.loading model
model = ETFC(50, 192, 21, 0.6, 1, 8)
model.load_state_dict(torch.load(dir))
# 2.predict
model.eval()
score_label = model(test, seq_length)
for i in range(len(score_label)):
for j in range(len(score_label[i])):
if score_label[i][j] < 0.5:
score_label[i][j] = 0
else:
score_label[i][j] = 1
return score_label
def pre_my(test_data, seq_length, output_path, names):
# models
h5_model = []
model_num = 10
for i in range(1, model_num + 1):
h5_model.append('model{}.h5'.format(str(i)))
# prediction
result = predict(test_data, seq_length, h5_model)
# label
peptides = ['AAP', 'ABP', 'ACP', 'ACVP', 'ADP', 'AEP', 'AFP', 'AHIVP', 'AHP', 'AIP', 'AMRSAP', 'APP', 'ATP',
'AVP',
'BBP', 'BIP',
'CPP', 'DPPIP',
'QSP', 'SBP', 'THP']
functions = []
for e in result:
temp = ''
for i in range(len(e)):
if e[i] == 1:
temp = temp + peptides[i] + ','
else:
continue
if temp == '':
temp = 'none'
if temp[-1] == ',':
temp = temp.rstrip(',')
functions.append(temp)
output_file = os.path.join(output_path, 'result.txt')
with open(output_file, 'w') as f:
for i in range(len(names)):
f.write(names[i])
f.write('functions:' + functions[i] + '\n')
if __name__ == '__main__':
args = ArgsGet()
file = args.file # fasta file
output_path = args.out_path # output path
# building output path directory
Path(output_path).mkdir(exist_ok=True)
# reading file and encoding
data, names, seq_length = get_data(file)
data = torch.LongTensor(data)
seq_length = torch.LongTensor(seq_length)
# prediction
pre_my(data, seq_length, output_path, names)