-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathSpatom.py
378 lines (352 loc) · 15.8 KB
/
Spatom.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
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
import torch
from Bio.PDB import *
import argparse
import os
from Bio.PDB import PDBParser, PDBIO, Select
import numpy as np
from Bio.PDB.DSSP import DSSP
from Bio import PDB
from torch_geometric.data import Data
import torch.utils.data
import sys
import freesasa
import math
from model.model import Spatom
from torch_geometric.loader import DataLoader
BATCH = 1
LEARN_RATE = 0.001
HIDDEN_DIM = 1024
LAYERS = 7
DROPOUT = 0.1
def logo():
print('\
* ____ _ *\n\
* / ___| _ __ __ _| |_ ___ _ __ ___ *\n\
* \___ \| \'_ \ / _` | __/ _ \| \'_ ` _ \ *\n\
* ___) | |_) | (_| | || (_) | | | | | | *\n\
* |____/| .__/ \__,_|\__\___/|_| |_| |_| *\n\
* |_| *')
def transpose(matrix):
new_matrix = []
for i in range(len(matrix[0])):
matrix_raw = []
for j in range(len(matrix)):
matrix_raw.append(matrix[j][i])
new_matrix.append(matrix_raw)
return new_matrix
def PSSM_file_to_PSSM_dict(protein,chains,PSSM):
Max_pssm = np.array([8., 9., 9., 9., 12., 9., 8., 8., 12., 9., 7., 8., 11.,
10., 9., 8., 8., 13., 10., 8.])
Min_pssm = np.array([-10., -11., -12., -12., -11., -10., -11., -11., -11., -10., -11.,
-11., -10., -11., -12., -11., -10., -11., -10., -11.])
PSSM_dict = {}
PSSM_matrix = []
for c in chains:
PSSM_file = open(PSSM + '/' + protein.upper() + '_' + c.upper() + '.PSSM', 'r')
for line in PSSM_file:
if line != None and len(line.split()) > 40:
PSSM_line = line.split()[2:22]
PSSM_line = list(map(float, PSSM_line))
PSSM_line = ((np.array(PSSM_line) - Min_pssm)/(Max_pssm - Min_pssm)).tolist()
PSSM_matrix.append(PSSM_line)
PSSM_dict[protein.lower()] = PSSM_matrix
return PSSM_dict
def PDB_proressing(real_PDB,protein):
class CA_Select(Select):
def accept_residue(self, residue):
return 1 if residue.id[0] == " " else 0
def accept_atom(self, atom):
if atom.get_name() == 'CA':
return True
else:
return False
class Residue_Select(Select):
def accept_residue(self, residue):
return 1 if residue.id[0] == " " else 0
pdb = PDBParser(QUIET=1).get_structure(protein, real_PDB)
io = PDBIO()
io.set_structure(pdb)
if not os.path.exists(sys.path[0]+"/CA_PDB"):
os.makedirs(sys.path[0]+"/CA_PDB")
io.save(sys.path[0]+'/CA_PDB/' + protein + ".pdb", CA_Select())
if not os.path.exists(sys.path[0]+"/Residue_PDB"):
os.makedirs(sys.path[0]+"/Residue_PDB")
io.save(sys.path[0]+'/Residue_PDB/' + protein + ".pdb", Residue_Select())
def Dist_adj(protein,chains):
def dictance(xyz, position):
xyz = xyz - xyz[position]
dictance = np.sqrt(xyz[:, 0] ** 2 + xyz[:, 1] ** 2 + xyz[:, 2] ** 2).tolist()
return dictance
p = PDBParser(QUIET=1)
structure = p.get_structure(protein, sys.path[0]+"/CA_PDB/" + protein + '.pdb')
distance_list = []
chain_list = []
for c in chains:
for residue in structure[0][c]:
for atom in residue:
chain_list.append(atom.get_vector().get_array())
for i, center in enumerate(chain_list):
distance_list.append(dictance(chain_list, i))
distance_dict = {}
distance_dict[protein.lower()] = np.array(distance_list)
return distance_dict
def seq_and_one_hot(protein, chains):
AA_index = {j: i for i, j in enumerate("ACDEFGHIKLMNPQRSTVWY")}
seq_dict = {}
one_hot_dict = {}
parser = PDBParser(QUIET=1)
structure = parser.get_structure(protein, sys.path[0] + '/Residue_PDB/' + protein + ".pdb")
ppb = PDB.PPBuilder()
one_hot_list = []
for c in chains:
model = structure[0][c.upper()]
seq = ''
for cc in ppb.build_peptides(model):
seq += cc.get_sequence()
seq_dict[(protein + c).lower()] = seq
for i in seq:
zero = [0] * 20
zero[AA_index[i]] = 1
one_hot_list.append(zero)
one_hot_dict[protein.lower()] = one_hot_list
return seq_dict, one_hot_dict
class ChianSelect(Select):
def __init__(self,chain_letter):
self.chain_letter = chain_letter
def accept_chain(self,chain):
if chain.get_id()==self.chain_letter:
return True
else:
return False
def get_RSA(protein,chains):
RSA_dict = {}
structure = freesasa.Structure(sys.path[0]+"/Residue_PDB/"+protein + '.pdb')
result = freesasa.calc(structure,freesasa.Parameters({'algorithm' : freesasa.LeeRichards,'n-slices' : 100,'probe-radius' : 1.4}))
residueAreas = result.residueAreas()
RSA = []
for c in chains:
for r in residueAreas[c.upper()].keys():
RSA_AA = []
RSA_AA.append(min(1,residueAreas[c.upper()][r].relativeTotal))
RSA_AA.append(min(1,residueAreas[c.upper()][r].relativePolar))
RSA_AA.append(min(1,residueAreas[c.upper()][r].relativeApolar))
RSA_AA.append(min(1,residueAreas[c.upper()][r].relativeMainChain))
if math.isnan(residueAreas[c.upper()][r].relativeSideChain):
RSA_AA.append(0)
else:
RSA_AA.append(min(1,residueAreas[c.upper()][r].relativeSideChain))
RSA.append(RSA_AA)
RSA_dict[protein.lower()] = RSA
return RSA_dict
def dssp_feature(protein,chains,fasta_dict):
SS_dict = {'H': 0, 'B': 1, 'E': 2, 'G': 3, 'I': 4, 'T': 5, 'S': 6, '-': 7}
p = PDBParser(QUIET=1)
pdb = p.get_structure(protein, sys.path[0]+'/Residue_PDB/' + protein + '.pdb')
io = PDBIO()
io.set_structure(pdb)
# io.save(sys.path[0]+'/Residue_PDB/'+protein + '_' + chain.upper() + '.pdb', ChianSelect(chain.upper()))
# structure = p.get_structure(protein,sys.path[0]+'/Residue_PDB/'+ protein + '_' + chain.upper() + '.pdb')
model = pdb[0]
dssp = DSSP(model, sys.path[0]+'/Residue_PDB/' + protein + '.pdb', dssp='mkdssp')
dssp_matrix_complex = []
for c in chains:
dssp_matrix = []
seq = ""
ref_seq = fasta_dict[(protein + c).lower()]
for i in dssp.keys():
if i[0] == c:
SS = dssp[i][2]
AA = dssp[i][1]
seq += AA
phi = dssp[i][4]
psi = dssp[i][5]
raw = []
raw.append(np.sin(phi * (np.pi / 180)))
raw.append(np.sin(psi * (np.pi / 180)))
raw.append(np.cos(phi * (np.pi / 180)))
raw.append(np.cos(psi * (np.pi / 180)))
ss_raw = [0] * 9
ss_raw[SS_dict[SS]] = 1
raw.extend(ss_raw)
dssp_matrix.append(raw)
pad = []
pad.append(np.sin(360 * (np.pi / 180)))
pad.append(np.sin(360 * (np.pi / 180)))
pad.append(np.cos(360 * (np.pi / 180)))
pad.append(np.cos(360 * (np.pi / 180)))
ss_pad = [0] * 9
ss_pad[-1] = 1
pad.extend(ss_pad)
pad_dssp_matrix = []
p_ref = 0
for i in range(len(seq)):
while p_ref < len(ref_seq) and seq[i] != ref_seq[p_ref]:
pad_dssp_matrix.append(pad)
p_ref += 1
if p_ref < len(ref_seq): # aa matched
pad_dssp_matrix.append(dssp_matrix[i])
p_ref += 1
if len(pad_dssp_matrix) != len(ref_seq):
for i in range(len(ref_seq) - len(pad_dssp_matrix)):
pad_dssp_matrix.append(pad)
dssp_matrix_complex.extend(pad_dssp_matrix)
dssp_dict = {}
dssp_dict[protein.lower()] = dssp_matrix_complex
return dssp_dict
def AA_property(protein,chains,seq_dict):
Side_Chain_Atom_num = {'A': 5.0, 'C': 6.0, 'D': 8.0, 'E': 9.0, 'F': 11.0, 'G': 4.0, 'H': 10.0, 'I': 8.0, 'K': 9.0,
'L': 8.0, 'M': 8.0, 'N': 8.0, 'P': 7.0, 'Q': 9.0, 'R': 11.0, 'S': 6.0, 'T': 7.0, 'V': 7.0,
'W': 14.0, 'Y': 12.0}
Side_Chain_Charge_num = {'A': 0.0, 'C': 0.0, 'D': -1.0, 'E': -1.0, 'F': 0.0, 'G': 0.0, 'H': 1.0, 'I': 0.0, 'K': 1.0,
'L': 0.0, 'M': 0.0, 'N': 0.0, 'P': 0.0, 'Q': 0.0, 'R': 1.0, 'S': 0.0, 'T': 0.0, 'V': 0.0,
'W': 0.0, 'Y': 0.0}
Side_Chain_hydrogen_bond_num = {'A': 2.0, 'C': 2.0, 'D': 4.0, 'E': 4.0, 'F': 2.0, 'G': 2.0, 'H': 4.0, 'I': 2.0,
'K': 2.0, 'L': 2.0, 'M': 2.0, 'N': 4.0, 'P': 2.0, 'Q': 4.0, 'R': 4.0, 'S': 4.0,
'T': 4.0, 'V': 2.0, 'W': 3.0, 'Y': 3.0}
Side_Chain_pKa = {'A': 7.0, 'C': 7.0, 'D': 3.65, 'E': 3.22, 'F': 7.0, 'G': 7.0, 'H': 6.0, 'I': 7.0, 'K': 10.53,
'L': 7.0, 'M': 7.0, 'N': 8.18, 'P': 7.0, 'Q': 7.0, 'R': 12.48, 'S': 7.0, 'T': 7.0, 'V': 7.0,
'W': 7.0, 'Y': 10.07}
Hydrophobicity = {'A': 1.8, 'C': 2.5, 'D': -3.5, 'E': -3.5, 'F': 2.8, 'G': -0.4, 'H': -3.2, 'I': 4.5, 'K': 3.9,
'L': 3.8, 'M': 1.9, 'N': -3.5, 'P': -1.6, 'Q': -3.5, 'R': -4.5, 'S': -0.8, 'T': -0.7, 'V': 4.2,
'W': -0.9, 'Y': -1.3}
AA_property_dict = {}
AA_protein = []
for c in chains:
for AA in seq_dict[(protein+c).lower()]:
AA_AA = []
AA_AA.append(Side_Chain_Atom_num[AA])
AA_AA.append(Side_Chain_Charge_num[AA])
AA_AA.append(Side_Chain_hydrogen_bond_num[AA])
AA_AA.append(Side_Chain_pKa[AA])
AA_AA.append(Hydrophobicity[AA])
AA_protein.append(AA_AA)
AA_property_dict[protein.lower()] = AA_protein
return AA_property_dict
def edge_weight(dist):
matrix = dist.clone()
softmax = torch.nn.Softmax(dim=0)
dist = softmax(1./(torch.log(torch.log(dist+2))))
dist[matrix>14] = 0
return dist
def feature_extract(protein,chains,PSSM_file):
PSSM_dict = PSSM_file_to_PSSM_dict(protein, chains, PSSM_file)
Dist_dict = Dist_adj(protein, chains)
fasta_dict, onehot_dict = seq_and_one_hot(protein, chains)
DSSP_dict = dssp_feature(protein, chains, fasta_dict)
RSA_dict = get_RSA(protein, chains)
AA_property_dict = AA_property(protein, chains, fasta_dict)
Datasets = []
feature = []
RSA = []
for k in range(len(onehot_dict[protein.lower()])):
AA = []
AA.extend(onehot_dict[protein.lower()][k])
AA.extend(PSSM_dict[protein.lower()][k])
AA.extend(RSA_dict[protein.lower()][k])
AA.extend(AA_property_dict[protein.lower()][k])
AA.extend(DSSP_dict[protein.lower()][k])
feature.append(AA)
RSA.append(RSA_dict[protein.lower()][k][0])
pos = np.where(np.array(RSA) >= 0.05)[0].tolist()
Dist = edge_weight(torch.tensor(Dist_dict[protein.lower()]))[pos, :][:, pos]
feature = torch.tensor(np.array(feature)[pos, :], dtype=torch.float)
adj = torch.tensor(np.where(np.array(Dist_dict[protein.lower()]) < 14, 1, 0)[pos, :][:, pos])
data = Data(x=feature)
data.dist = Dist
data.POS = pos
length = len(onehot_dict[protein.lower()])
data.length = length
data.adj = adj
Datasets.append(data)
return Datasets
def predict(protein,chains,pdb_file,output_path,PSSM):
def test(model,test_set,output_path):
test_loader = DataLoader(test_set, batch_size=1)
model.eval()
all_pred = []
with torch.no_grad():
for step, data in enumerate(test_loader):
feature = torch.autograd.Variable(data.x.to(DEVICE, dtype=torch.float))
dist = torch.autograd.Variable(data.dist.to(DEVICE, dtype=torch.float))
adj = torch.autograd.Variable(data.adj.to(DEVICE, dtype=torch.float))
pos = data.POS[0]
length = data.length.item()
pred = model(feature, dist, adj)
pred = pred.cpu().numpy().tolist()
predict_protein = [0] * length
for k, i in enumerate(pos):
predict_protein[i] = pred[k]
all_pred.extend(predict_protein)
result = np.where(np.array(all_pred) > 0.27, 1, 0).tolist()
protein = args.protein
i = 0
f = open(output_path + 'predict_' + protein + '.txt', 'w')
f.write('Protein: ' + protein + '\n')
f.write('Number ' + ' Chain' + ' Amino_Acid ' + ' Predict ' + ' Score' + '\n')
for c in chains:
parser = PDBParser(QUIET=1)
structure = parser.get_structure(protein, sys.path[0] + '/Residue_PDB/' + protein + '.pdb')
residue_id_list = []
for aa in structure[0][c.upper()].get_residues():
residue_id_list.append(aa.get_id()[1])
for k, p in enumerate(fasta_dict[(protein + c).lower()]):
f.write(' ' + "{:>3d}".format(residue_id_list[k]) + ' ' + c + ' ' + p + ' ' + str(result[i])
+ ' ' + "{:.3f}".format(all_pred[i]) + '\n')
i += 1
f.close()
logo()
if not os.path.exists(output_path):
os.makedirs(output_path)
DEVICE = torch.device("cuda" if torch.cuda.is_available() else "cpu")
print('Device: ',DEVICE)
best_model = Spatom().to(DEVICE)
best_model.load_state_dict(torch.load(sys.path[0]+'/model/best_model.dat'))
test_set = feature_extract(protein,chains, PSSM)
fasta_dict, _ = seq_and_one_hot(protein, chains)
test(best_model,test_set,output_path)
print('Done!')
return 1
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument('--protein', '-p', help='protein name', type = str)
parser.add_argument('--PSSM', '-m', help='path_to_PSSM_folder, save the PSSM files of corresponding chains into this folder', type=str)
parser.add_argument('--pdb', '-b', help='path_to_pdb/pdb_file', type=str)
parser.add_argument('--output', '-o', help='path_to_output',default=sys.path[0]+'/result/', type=str)
args = parser.parse_args()
if args.PSSM:
fileList = os.listdir(args.PSSM)
for file in fileList:
os.rename(args.PSSM + '/' + file, args.PSSM + '/' + file.upper())
if args.protein == None or args.PSSM == None:
print('Please check your input!')
sys.exit('error!')
else:
if args.pdb == None:
print('Down load PDB for input protein name!')
if not os.path.exists(sys.path[0]+'/real_PDB'):
os.makedirs(sys.path[0]+'/real_PDB')
pdb = PDBList()
pdb.retrieve_pdb_file(args.protein, pdir=sys.path[0]+'/real_PDB', file_format='pdb')
PDB_proressing(sys.path[0]+'/real_PDB/pdb'+args.protein.lower()+'.ent', args.protein)
pre_pdb = PDBParser(QUIET=1).get_structure(args.protein, sys.path[0]+"/Residue_PDB/"+args.protein + '.pdb').get_chains()
chains = []
for _ in pre_pdb:
chain = _.get_id()
chains.append(chain)
if not os.path.isfile(args.PSSM + '/' + args.protein.upper() + '_' + chain.upper() + '.PSSM'):
print('Please check the pssm file!')
sys.exit('error!')
predict(args.protein,chains,sys.path[0]+'/real_PDB/pdb'+args.protein.lower()+'.ent',args.output,args.PSSM)
else:
print('Using given PDB files!')
PDB_proressing(args.pdb, args.protein)
pre_pdb = PDBParser(QUIET=1).get_structure(args.protein, sys.path[0]+"/Residue_PDB/"+args.protein + '.pdb').get_chains()
chains = []
for _ in pre_pdb:
chain = _.get_id()
chains.append(chain)
print(chain)
if not os.path.isfile(args.PSSM + '/' + args.protein.upper() + '_' + chain.upper() + '.PSSM'):
print('Please check the pssm file!')
sys.exit('error!')
predict(args.protein,chains,args.pdb,args.output,args.PSSM)