-
Notifications
You must be signed in to change notification settings - Fork 0
/
optimize_property_chemspace.py
849 lines (733 loc) · 40.5 KB
/
optimize_property_chemspace.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
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
import argparse
import os
import sys
# for linux env.
sys.path.insert(0,'..')
from distutils.util import strtobool
import torch
import numpy as np
import pandas as pd
import torch
from rdkit import Chem, DataStructs
from rdkit.Chem import AllChem, Draw
from tdc import Oracle
from sklearn.preprocessing import scale, StandardScaler
from data.data_loader import NumpyTupleDataset
from data import transform_qm9, transform_zinc250k
from data.transform_zinc250k import zinc250_atomic_num_list, transform_fn_zinc250k
# from mflow.generate import generate_mols_along_axis
from mflow.models.hyperparams import Hyperparameters
from mflow.models.utils import check_validity, adj_to_smiles, _to_numpy_array
from mflow.utils.model_utils import load_model, smiles_to_adj
from mflow.models.model import rescale_adj
import mflow.utils.environment as env
import time
import functools
print = functools.partial(print, flush=True)
GSK3B_scorer = Oracle(name = 'GSK3B')
DRD2_scorer = Oracle(name = 'DRD2')
JNK3_scorer = Oracle(name = 'JNK3')
# Define helper oracle functions that take in a RDKit Mol object and return the specfied molecular property
def check_DRD2(gen_smiles):
score = DRD2_scorer(Chem.MolToSmiles(gen_smiles))
return score
def check_JNK3(gen_smiles):
score = JNK3_scorer(Chem.MolToSmiles(gen_smiles))
return score
def check_GSK3B(gen_smiles):
score = GSK3B_scorer(Chem.MolToSmiles(gen_smiles))
return score
def get_z(model, mol, device):
"""
Convert a given SMILES string into its corresponding molecular graph representation and use a trained model to
extract a latent representation vector from the molecular graph. The returned representation vector is in the form
of a NumPy array.
Args:
model: An instance of a trained PyTorch model
mol (str): A SMILES string
device (torch.device, optional): A PyTorch device instance to be used for running the computation. Default is None,
indicating that the computation will be performed on the CPU.
Returns:
z (ndarray): A NumPy array representing the molecular latent representation extracted from the given molecular
structure.
"""
adj_idx, x_idx = smiles_to_adj(mol, data_name='zinc250k')
if device:
adj_idx = adj_idx.to(device)
x_idx = x_idx.to(device)
adj_normalized = rescale_adj(adj_idx).to(device)
z_idx, _ = model(adj_idx, x_idx, adj_normalized)
z_idx[0] = z_idx[0].reshape(z_idx[0].shape[0], -1)
z_idx[1] = z_idx[1].reshape(z_idx[1].shape[0], -1)
z_idx = torch.cat((z_idx[0], z_idx[1]), dim=1).squeeze(dim=0) # h:(1,45), adj:(1,324) -> (1, 369) -> (369,)
z = np.expand_dims(_to_numpy_array(z_idx), axis=0)
return z
def optimize_mol(model, smiles, direction, num_range, path_len=21, sim_cutoff=0,
data_name='zinc250k', atomic_num_list=zinc250_atomic_num_list, property_name='plogp', device=None):
"""
Optimize a given molecular structure for a single property.
Args:
model: An instance of a trained PyTorch model to be used for molecular optimization in latent space.
smiles (str): A SMILES string representing the molecular structure to be optimized.
direction (ndarray): A NumPy array specifying the direction in the latent space to optimize the molecule towards.
The array has the same shape as the latent representation vector.
num_range (float): A scalar value specifying the range of the optimization path in the given direction.
path_len (int, optional): An integer value specifying the number of points on the optimization path. Default is 21.
sim_cutoff (float, optional): A scalar value specifying the Tanimoto similarity threshold for accepting a new
molecular structure. Default is 0.
data_name (str, optional): The name of the dataset used to train the model. Defaults to 'zinc250k'.
atomic_num_list (list, optional): A list of atomic numbers to be used for molecular graph construction. Default is
'zinc250_atomic_num_list'.
property_name (str, optional): A string specifying the molecular property name to be optimized. Default is 'plogp'.
device (torch.device, optional): The device to use for processing. Default is None, indicating that the computation
will be performed on the CPU.
Returns:
smiles_path (list): A list of SMILES strings representing the molecular structures along the optimization path.
results (list): A list of tuples containing the SMILES string, the corresponding property value, the Tanimoto
similarity to the input molecule, and the input molecule SMILES string for each valid molecule
generated during the optimization process.
start (tuple): A tuple containing the input molecule SMILES string, its corresponding property value, and None.
"""
if property_name == 'qed':
propf = env.qed # [0,1]
elif property_name == 'plogp':
propf = env.penalized_logp # unbounded, normalized later???
elif property_name == 'drd2':
propf = check_DRD2
elif property_name == 'jnk3':
propf = check_JNK3
elif property_name == 'gsk3b':
propf = check_GSK3B
else:
raise ValueError("Wrong property_name{}".format(property_name))
model.eval()
mol = Chem.MolFromSmiles(smiles)
fp1 = AllChem.GetMorganFingerprint(mol, 2)
start = (smiles, propf(mol), None) # , mol)
z = get_z(model, smiles, device)
with torch.no_grad():
distances = np.linspace(-num_range,num_range,path_len)
distances = distances.tolist()
print ('Z Shape: ', z.shape, 'Direction Shape: ', direction.shape)
z_to_decode = []
for j in range(len(distances)):
z_to_decode.append(z + distances[j]*direction)
z_to_decode = torch.from_numpy(np.array(z_to_decode)).squeeze().float().to(device)
adj, x = model.reverse(z_to_decode)
smiles_path = adj_to_smiles(adj.cpu(), x.cpu(), atomic_num_list)
val_res = check_validity(adj, x, atomic_num_list)
valid_mols = val_res['valid_mols']
valid_smiles = val_res['valid_smiles']
results = []
sm_set = set()
sm_set.add(smiles)
for m, s in zip(valid_mols, valid_smiles):
if s in sm_set:
continue
sm_set.add(s)
p = propf(m)
fp2 = AllChem.GetMorganFingerprint(m, 2)
sim = DataStructs.TanimotoSimilarity(fp1, fp2)
if sim >= sim_cutoff:
results.append((s, p, sim, smiles))
# smile, property, similarity, mol
results.sort(key=lambda tup: tup[1], reverse=True)
return smiles_path, results, start
def optimize_mol_multi_prop(model, smiles, directions, num_range, path_len=21, sim_cutoff=0,
data_name='zinc250k', atomic_num_list=zinc250_atomic_num_list, device=None):
"""
Optimize a given molecular structure for multiple properties.
Args:
model (nn.Module): An instance of a trained PyTorch model to be used for molecular optimization in latent space.
smiles (str): A SMILES string representing the molecular structure to be optimized.
directions (list of numpy.ndarray): A list of direction vectors corresponding to the properties to be optimized.
num_range (float): A scalar value specifying the range of the optimization path in the given directions.
path_len (int, optional): The number of intermediate points on the optimization path. Defaults to 21.
sim_cutoff (float, optional): A scalar value specifying the Tanimoto similarity threshold for accepting a new
molecular structure. Defaults to 0.
data_name (str, optional): The name of the dataset used to train the model. Defaults to 'zinc250k'.
atomic_num_list (list of int, optional): A list of atomic numbers for the atoms present in the molecular structure. Defaults to zinc250_atomic_num_list.
device (torch.device, optional): The device to use for processing. Default is None, indicating that the computation will be performed on the CPU.
Returns:
smiles_path (list of str): A list of SMILES strings representing the optimized molecular structures at each intermediate point.
results_1 (list of tuple): A list of tuples, where each tuple contains the SMILES string, the property value and the Tanimoto similarity of the optimized molecular structures with respect to the first property.
results_2 (list of tuple): A list of tuples, where each tuple contains the SMILES string, the property value and the Tanimoto similarity of the optimized molecular structures with respect to the second property.
start_1 (tuple): A tuple containing the SMILES string, the property value and None for the starting molecular structure with respect to the first property.
start_2 (tuple): A tuple containing the SMILES string, the property value and None for the starting molecular structure with respect to the second property.
"""
propf_1 = env.qed # [0,1]
propf_2 = env.penalized_logp # unbounded, normalized later???
if len(directions) > 1:
combined_directions = directions[0] * directions[1]
pos_attributes = (combined_directions >= 0) * directions[1]
direction = directions[0] + pos_attributes
else:
direction = directions[0]
model.eval()
mol = Chem.MolFromSmiles(smiles)
fp1 = AllChem.GetMorganFingerprint(mol, 2)
start_1 = (smiles, propf_1(mol), None) # , mol)
start_2 = (smiles, propf_2(mol), None) # , mol)
z = get_z(model, smiles, device)
with torch.no_grad():
distances = np.linspace(-num_range,num_range,path_len)
distances = distances.tolist()
print ('Z Shape: ', z.shape, 'Direction Shape: ', direction.shape)
z_to_decode = []
for j in range(len(distances)):
z_to_decode.append(z + distances[j]*direction)
z_to_decode = torch.from_numpy(np.array(z_to_decode)).squeeze().float().to(device)
adj, x = model.reverse(z_to_decode)
smiles_path = adj_to_smiles(adj.cpu(), x.cpu(), atomic_num_list)
val_res = check_validity(adj, x, atomic_num_list)
valid_mols = val_res['valid_mols']
valid_smiles = val_res['valid_smiles']
results_1 = []
results_2 = []
sm_set = set()
sm_set.add(smiles)
for m, s in zip(valid_mols, valid_smiles):
if s in sm_set:
continue
sm_set.add(s)
p1 = propf_1(m)
p2 = propf_2(m)
fp2 = AllChem.GetMorganFingerprint(m, 2)
sim = DataStructs.TanimotoSimilarity(fp1, fp2)
if sim >= sim_cutoff:
results_1.append((s, p1, sim, smiles))
results_2.append((s, p2, sim, smiles))
return smiles_path, results_1, results_2, start_1, start_2
def load_property_csv(data_name):
"""
Load property values from a CSV file for a given dataset.
Args:
data_name (str): Name of the dataset, either 'qm9' or 'zinc250k'.
Returns:
tuples (list): A list of tuples, each containing property values for a single molecule.
prop_to_idx (dict): A dictionary that maps property names to their corresponding index in the tuple.
"""
if data_name == 'qm9':
# Total: 133885
filename = 'data/qm9_properties.csv'
elif data_name == 'zinc250k':
# Total: 249455
filename = 'data/zinc250k_properties.csv'
df = pd.read_csv(filename) # smile,qed,plogp,MolLogP,MolWt,sa,drd2,jnk3,gsk3b
tuples = [tuple(x) for x in df.values]
props = ['smile','qed','plogp','MolLogP','MolWt','sa','drd2','jnk3','gsk3b']
values = np.linspace(0,8,9, dtype=int)
prop_to_idx = dict(zip(props, values))
print('Load {} done, length: {}'.format(filename, len(tuples)))
return tuples, prop_to_idx
def find_top_score_smiles(model, device, data_name, property_name, prop_to_idx, train_prop, topk, atomic_num_list, path):
"""
Find top k optimized molecules for a given property.
Args:
model (object): A PyTorch model for optimizing molecules.
device (str): The device to use for optimization.
data_name (str): The name of the dataset.
property_name (str): The name of the property score to optimize for.
prop_to_idx (dict): A dictionary that maps property names to their corresponding column index in the dataset.
train_prop (list): A list of tuples containing the molecular properties of the training dataset.
topk (int): The number of optimized molecules to return.
atomic_num_list (list): A list of atomic numbers to consider for the optimized molecules.
path (str): The path to save the optimized molecules.
Returns:
Saves a list of tuples containing the optimized molecules, their corresponding property score,
their similarity to the reference molecule and the reference molecule itself to a csv at the given path
"""
start_time = time.time()
idx = prop_to_idx[property_name] # smile,qed,plogp,MolLogP,MolWt,sa,drd2,jnk3,gsk3b
print('Finding top {} score'.format(property_name))
train_prop_sorted = sorted(train_prop, key=lambda tup: tup[idx], reverse=True)
if not os.path.exists('./'+data_name+'_chemspace_opt/'+f'{path}'):
os.makedirs('./'+data_name+'_chemspace_opt/'+f'{path}')
direction = np.load('./boundaries_'+args.data_name+'/boundary_'+property_name+'.npy')
result_list = []
for i, r in enumerate(train_prop_sorted):
if i >= topk:
break
if i % 50 == 0:
print('Optimization {}/{}, time: {:.2f} seconds'.format(i, topk, time.time() - start_time))
smile = r[prop_to_idx['smile']]
_ , results, _ = optimize_mol(model, smile, direction, num_range=100, path_len=21, sim_cutoff=0.0,
data_name=data_name, atomic_num_list=atomic_num_list,
property_name=property_name, device=device)
result_list.extend(results) # results: [(smile2, property, sim, smile1), ...]
result_list.sort(key=lambda tup: tup[1], reverse=True)
# check novelty
train_smile = set()
for i, r in enumerate(train_prop_sorted):
smile = r[prop_to_idx['smile']]
train_smile.add(smile)
mol = Chem.MolFromSmiles(smile)
smile2 = Chem.MolToSmiles(mol, isomericSmiles=True)
train_smile.add(smile2)
result_list_novel = []
start_time = time.time()
idx = prop_to_idx[property_name] # smile,qed,plogp,MolLogP,MolWt,sa,drd2,jnk3,gsk3b
print('Finding top {} score'.format(property_name))
train_prop_sorted = sorted(train_prop, key=lambda tup: tup[idx], reverse=True)
if not os.path.exists('./'+data_name+'_chemspace_opt/'+f'{path}'):
os.makedirs('./'+data_name+'_chemspace_opt/'+f'{path}')
direction = np.load('./boundaries_'+args.data_name+'/boundary_'+property_name+'.npy')
result_list = []
for i, r in enumerate(train_prop_sorted):
if i >= topk:
break
if i % 50 == 0:
print('Optimization {}/{}, time: {:.2f} seconds'.format(i, topk, time.time() - start_time))
smile = r[prop_to_idx['smile']]
_ , results, _ = optimize_mol(model, smile, direction, num_range=100, path_len=21, sim_cutoff=0.0,
data_name=data_name, atomic_num_list=atomic_num_list,
property_name=property_name, device=device)
result_list.extend(results) # results: [(smile2, property, sim, smile1), ...]
result_list.sort(key=lambda tup: tup[1], reverse=True)
# check novelty
train_smile = set()
for i, r in enumerate(train_prop_sorted):
smile = r[prop_to_idx['smile']]
train_smile.add(smile)
mol = Chem.MolFromSmiles(smile)
smile2 = Chem.MolToSmiles(mol, isomericSmiles=True)
train_smile.add(smile2)
result_list_novel = []
for i, r in enumerate(result_list):
smile, score, sim, smile_original = r
if smile not in train_smile:
result_list_novel.append(r)
# dump results
f = open('./'+data_name+'_chemspace_opt/'+f'{path}' + '_discovered_sorted.csv', "w")
for r in result_list_novel:
smile, score, sim, smile_original = r
f.write('{},{},{},{}\n'.format(score, smile, sim, smile_original))
f.flush()
f.close()
print('Dump done!')
def find_top_score_smiles_multi_prop(model, device, data_name, property_name, prop_to_idx, train_prop, topk, atomic_num_list, path):
"""
Find top k optimized molecules for multiple properties.
Args:
model (object): A PyTorch model for optimizing molecules.
device (str): The device to use for optimization.
data_name (str): The name of the dataset.
property_name (str): The name of the property score to optimize for.
prop_to_idx (dict): A dictionary that maps property names to their corresponding column index in the dataset.
train_prop (list): A list of tuples containing the molecular properties of the training dataset.
topk (int): The number of optimized molecules to return.
atomic_num_list (list): A list of atomic numbers to consider for the optimized molecules.
path (str): The path to save the optimized molecules.
Returns:
Saves a list of tuples containing the optimized molecules, their corresponding property scores,
their similarity to the reference molecule, the reference molecule itself, the change in each property,
and the combined improvement to a csv at the given path
"""
start_time = time.time()
print('Finding top {} score'.format(property_name))
train_list = list(zip(*train_prop))
qed_scaler = StandardScaler()
plogp_scaler = StandardScaler()
qed_list = np.array(train_list[prop_to_idx['qed']]).reshape(-1,1)
plogp_list = np.array(train_list[prop_to_idx['plogp']]).reshape(-1,1)
qed_scaler.fit(qed_list)
plogp_scaler.fit(plogp_list)
print(qed_scaler.mean_)
print(plogp_scaler.mean_)
qed_scaled = qed_scaler.transform(qed_list).reshape(-1)
plogp_scaled = plogp_scaler.transform(plogp_list).reshape(-1)
train_prop_extended = []
for i, smi in enumerate(train_prop):
scaled_prop = (qed_scaled[i]+plogp_scaled[i],)
new_tup = smi + scaled_prop
train_prop_extended.append(new_tup)
print('Finding top qed_plogp score')
train_prop_sorted = sorted(train_prop_extended, key=lambda tup: tup[-1], reverse=True)
if not os.path.exists('./'+data_name+'_chemspace_opt/'+f'{path}'):
os.makedirs('./'+data_name+'_chemspace_opt/'+f'{path}')
directions = []
props = ['qed', 'plogp']
for prop_name in props:
direction = np.load('./boundaries_'+args.data_name+'/boundary_'+prop_name+'.npy')
directions.append(direction)
result_list = []
for i, r in enumerate(train_prop_sorted):
if i >= topk:
break
if i % 50 == 0:
print('Optimization {}/{}, time: {:.2f} seconds'.format(i, topk, time.time() - start_time))
smile = r[prop_to_idx['smile']]
qed = r[prop_to_idx['qed']]
plogp = r[prop_to_idx['plogp']]
_ , results1, results2, _ , _ = optimize_mol_multi_prop(model, smile, directions, num_range=100, path_len=21, sim_cutoff=0.0,
data_name=data_name, atomic_num_list=atomic_num_list, device=device)
qed_results = np.array(list(zip(*results1))[1]).reshape(-1,1)
plogp_results = np.array(list(zip(*results2))[1]).reshape(-1,1)
improvement_qed = qed_scaler.transform(qed_results).reshape(-1)
improvement_plogp = plogp_scaler.transform(plogp_results).reshape(-1)
improvement_combined = improvement_qed + improvement_plogp
results_1_sorted = [r for _,r in sorted(zip(improvement_combined, results1), reverse=True)]
results_2_sorted = [r for _,r in sorted(zip(improvement_combined, results2), reverse=True)]
improvement_combined = np.sort(improvement_combined)[::-1]
for r1, r2, imp in zip(results_1_sorted, results_2_sorted, improvement_combined):
smile1, property1, sim, _ = r1
smile2, property2, _ , _ = r2
qed_delta = property1 - qed
plogp_delta = property2 - plogp
result_list.append((smile1, smile, sim, property1, property2, qed, plogp, qed_delta, plogp_delta, imp))
result_list.sort(key=lambda tup: tup[-1], reverse=True)
# check novelty
train_smile = set()
for i, r in enumerate(train_prop_sorted):
qed, plogp, smile, _ = r
train_smile.add(smile)
mol = Chem.MolFromSmiles(smile)
smile2 = Chem.MolToSmiles(mol, isomericSmiles=True)
train_smile.add(smile2)
result_list_novel = []
for i, r in enumerate(result_list):
smile = r[0]
if smile not in train_smile:
result_list_novel.append(r)
# dump results
f = open('./'+data_name+'_chemspace_opt/'+f'{path}' + '_discovered_sorted.csv', "w")
for r in result_list_novel:
smile, smile_original, sim, qed, plogp, qed_original, plogp_original, qed_delta, plogp_delta, imp = r
f.write('{},{},{},{},{},{},{},{},{},{}\n'.format(imp, qed, plogp, smile, smile_original, sim, qed_original, plogp_original, qed_delta, plogp_delta))
f.flush()
f.close()
print('Dump done!')
def constrain_optimization_smiles(model, device, data_name, property_name, prop_to_idx, train_prop, topk,
atomic_num_list, path, path_range=100, sim_cutoff=0.0):
"""
Optimize molecules for a given property, while maintaining the desired similarity (Tanimoto) to
the given reference molecule.
Args:
model (object): A PyTorch model for optimizing molecules.
device (str): The device to use for optimization.
data_name (str): The name of the dataset.
property_name (str): The name of the property score to optimize for.
prop_to_idx (dict): A dictionary that maps property names to their corresponding column index in the dataset.
train_prop (list): A list of tuples containing the molecular properties of the training dataset.
topk (int): The number of optimized molecules to return.
atomic_num_list (list): A list of atomic numbers to consider for the optimized molecules.
path (str): The path to save the optimized molecules.
path_range (int, optional): The maximum number of steps that can be taken in each direction during optimization.
Defaults to 100.
sim_cutoff (float, optional): The similarity cutoff used to filter out similar molecules during optimization.
Defaults to 0.0.
Returns:
Saves a list of tuples containing the optimized molecules, their corresponding property score,
their similarity to the reference molecule, the reference molecule, the original property score,
and the change in property to a csv at the given path
"""
start_time = time.time()
idx = prop_to_idx[property_name] # smile,qed,plogp,MolLogP,MolWt,sa,drd2,jnk3,gsk3b
print('Constrained optimization of {} score'.format(property_name))
train_prop_sorted = sorted(train_prop, key=lambda tup: tup[idx]) #, reverse=True) # qed, plogp, smile
result_list = []
nfail = 0
if not os.path.exists('./'+data_name+'_chemspace_consopt/'+f'{path}'):
os.makedirs('./'+data_name+'_chemspace_consopt/'+f'{path}')
direction = np.load('./boundaries_'+data_name+'/boundary_'+property_name+'.npy')
for i, r in enumerate(train_prop_sorted):
if i >= topk:
break
if i % 50 == 0:
print('Optimization {}/{}, time: {:.2f} seconds'.format(i, topk, time.time() - start_time))
smile = r[prop_to_idx['smile']]
prop = r[idx]
_ , results, _ = optimize_mol(model, smile, direction, num_range=path_range, path_len=21, sim_cutoff=sim_cutoff,
data_name=data_name, atomic_num_list=atomic_num_list,
property_name=property_name, device=device)
# for idx, smi_save in enumerate(smiles_path):
# np.save(open(filepath+'_'+str(i)+'_'+str(idx)+'.npy','wb'),smi_save)
if len(results) > 0:
smile2, property2, sim, _ = results[0]
prop_delta = property2 - prop
if prop_delta >= 0:
result_list.append((smile2, property2, sim, smile, prop, prop_delta))
else:
nfail += 1
print('Failure:{}:{}'.format(i, smile))
else:
nfail += 1
print('Failure:{}:{}'.format(i, smile))
print(result_list)
df = pd.DataFrame(result_list,
columns=['smile_new', 'prop_new', 'sim', 'smile_old', 'prop_old', 'prop_delta'])
print(df.describe())
df.to_csv('./'+data_name+'_chemspace_consopt/'+f'{path}'+'_constrain_optimization.csv', index=False)
print('Dump done!')
print('nfail:{} in total:{}'.format(nfail, topk))
print('success rate: {}'.format((topk-nfail)*1.0/topk))
def constrain_optimization_smiles_multi_prop(model, device, data_name, property_name, prop_to_idx, train_prop, topk,
atomic_num_list, path, path_range=100, sim_cutoff=0.0):
"""
Optimize molecules for multiple properties, while maintaining the desired similarity (Tanimoto) to
the given reference molecule.
Args:
model (object): A PyTorch model for optimizing molecules.
device (str): The device to use for optimization.
data_name (str): The name of the dataset.
property_name (str): The name of the property score to optimize for.
prop_to_idx (dict): A dictionary that maps property names to their corresponding column index in the dataset.
train_prop (list): A list of tuples containing the molecular properties of the training dataset.
topk (int): The number of optimized molecules to return.
atomic_num_list (list): A list of atomic numbers to consider for the optimized molecules.
path (str): The path to save the optimized molecules.
path_range (int, optional): The maximum number of steps that can be taken in each direction during optimization.
Defaults to 100.
sim_cutoff (float, optional): The similarity cutoff used to filter out similar molecules during optimization.
Defaults to 0.0.
Returns:
Saves a list of tuples containing the optimized molecules, their corresponding property scores,
their similarity to the reference molecule, the reference molecule, the original property scores,
and the change in properties to a csv at the given path
"""
start_time = time.time()
train_list = list(zip(*train_prop))
qed_scaled = scale(train_list[prop_to_idx['qed']])
plogp_scaled = scale(train_list[prop_to_idx['plogp']])
train_prop_extended = []
for i, smi in enumerate(train_prop):
scaled_prop = (qed_scaled[i]+plogp_scaled[i],)
new_tup = smi + scaled_prop
train_prop_extended.append(new_tup)
print('Constrained optimization of qed_plogp score'.format(property_name))
train_prop_sorted = sorted(train_prop_extended, key=lambda tup: tup[-1]) #, reverse=True) # qed, plogp, smile
result_list = []
nfail = 0
count = 0
filepath = './'+data_name+'_chemspace_consopt/'+f'{path}'+'/smiles'
if not os.path.exists('./'+data_name+'_chemspace_consopt/'+f'{path}'):
os.makedirs('./'+data_name+'_chemspace_consopt/'+f'{path}')
directions = []
props = ['qed', 'plogp']
for prop_name in props:
direction = np.load('./boundaries_'+args.data_name+'/boundary_'+prop_name+'.npy')
directions.append(direction)
for i, r in enumerate(train_prop_sorted):
if i >= topk:
break
if i % 50 == 0:
print('Optimization {}/{}, time: {:.2f} seconds'.format(i, topk, time.time() - start_time))
smile = r[prop_to_idx['smile']]
qed = r[prop_to_idx['qed']]
plogp = r[prop_to_idx['plogp']]
smiles_path, results_1, results_2, _, _ = optimize_mol_multi_prop(model, smile, directions, num_range=path_range, path_len=21, sim_cutoff=sim_cutoff,
data_name=data_name, atomic_num_list=atomic_num_list, device=device)
for idx, smi_save in enumerate(smiles_path):
np.save(open(filepath+'_'+str(i)+'_'+str(idx)+'.npy','wb'),smi_save)
if len(results_1) > 0 and len(results_2) > 0:
improvement_1 = [(result[1] - qed)/qed for result in results_1]
improvement_2 = [(result[1] - plogp)/plogp for result in results_2]
improvement_combined = improvement_1 + improvement_2
results_1_sorted = [r for _,r in sorted(zip(improvement_combined, results_1), reverse=True)]
results_2_sorted = [r for _,r in sorted(zip(improvement_combined, results_2), reverse=True)]
smile1, property1, sim, _ = results_1_sorted[0]
smile2, property2, _ , _ = results_2_sorted[0]
qed_delta = property1 - qed
plogp_delta = property2 - plogp
if qed_delta >=0 and plogp_delta >= 0:
result_list.append((smile1, smile, sim, property1, property2, qed, plogp, qed_delta, plogp_delta))
else:
nfail += 1
print('Failure:{}:{}'.format(i, smile))
else:
count += 1
nfail += 1
print('Failure:{}:{}'.format(i, smile))
df = pd.DataFrame(result_list,
columns=['smile_new', 'smile_old', 'sim', 'qed_new', 'plogp_new', 'qed_old', 'plogp_old', 'qed_delta', 'plogp_delta'])
print(df.describe())
df.to_csv('./'+data_name+'_chemspace_consopt/'+f'{path}'+'_constrain_optimization_multi.csv', index=False)
print('Dump done!')
print('nfail:{} in total:{}'.format(nfail, topk))
print('success rate: {}'.format((topk-nfail)*1.0/topk))
print(count)
def plot_top_qed_mol():
"""
Plot the top 25 molecules sorted by property score.
"""
import cairosvg
filename = 'qed_discovered_sorted.csv'
df = pd.read_csv(filename, header=None, names=['qed', 'Smile', 'Similarity', 'Smile Original'])
vmol = []
vlabel = []
for index, row in df.head(n=25).iterrows():
score, smile, sim, smile_old = row
print(score)
vmol.append(Chem.MolFromSmiles(smile))
vlabel.append('{:.3f}'.format(score))
svg = Draw.MolsToGridImage(vmol, legends=vlabel, molsPerRow=5, #5,
subImgSize=(120, 120), useSVG=True) # , useSVG=True
cairosvg.svg2pdf(bytestring=svg.encode('utf-8'), write_to="top_qed_moflow.pdf")
cairosvg.svg2png(bytestring=svg.encode('utf-8'), write_to="top_qed_moflow.png")
# print('Dump {}.png/pdf done'.format(filepath))
img = Draw.MolsToGridImage(vmol, legends=vlabel, molsPerRow=5,
subImgSize=(300, 300), useSVG=True)
# print(img)
def plot_mol_constraint_opt():
"""
Plot the molecular structures of two given SMILES strings along with their corresponding property values.
"""
import cairosvg
vsmiles = ['O=C(NCc1ccc2c3c(cccc13)C(=O)N2)c1ccc(F)cc1',
'O=C(NCC1=Cc2c[nH]c(=O)c3cccc1c23)c1ccc(F)cc1']
vmol = [Chem.MolFromSmiles(s) for s in vsmiles]
vplogp = ['{:.2f}'.format(env.penalized_logp(mol)) for mol in vmol]
# vhighlight = [vmol[0].GetSubstructMatch(Chem.MolFromSmiles('C2=C1C=CC=C3C1=C(C=C2)NC3')),
# vmol[1].GetSubstructMatch(Chem.MolFromSmiles('C4=CC6=C5C4=CC=CC5=C[N](=C6)[H]'))]
svg = Draw.MolsToGridImage(vmol, legends=vplogp, molsPerRow=2,
subImgSize=(250, 100), useSVG=True)
#highlightAtoms=vhighlight) # , useSVG=True
cairosvg.svg2pdf(bytestring=svg.encode('utf-8'), write_to="copt2.pdf")
cairosvg.svg2png(bytestring=svg.encode('utf-8'), write_to="copt2.png")
def plot_mol_matrix():
"""
Plot a matrix representation of the molecular structure and save the images as PDF and PNG files.
"""
import cairosvg
import seaborn as sns
import matplotlib.pyplot as plt
smiles = 'CN(C)C(=N)NC(=N)N' #'CC(C)NC1=CC=CO1' #'CC1=C(SC(=C1)C(=O)NCC2=NOC=C2)Br'
bond, atoms = smiles_to_adj(smiles, 'qm9')
bond = bond[0]
atoms = atoms[0]
# def save_mol_png(mol, filepath, size=(100, 100)):
# Draw.MolToFile(mol, filepath, size=size)
Draw.MolToImageFile(Chem.MolFromSmiles(smiles), 'mol.pdf')
# save_mol_png(Chem.MolFromSmiles(smiles), 'mol.png')
svg = Draw.MolsToGridImage([Chem.MolFromSmiles(smiles)], legends=[], molsPerRow=1,
subImgSize=(250, 250), useSVG=True)
# highlightAtoms=vhighlight) # , useSVG=True
cairosvg.svg2pdf(bytestring=svg.encode('utf-8'), write_to="mol.pdf")
cairosvg.svg2png(bytestring=svg.encode('utf-8'), write_to="mol.png")
# sns.set()
# ax = sns.heatmap(1-atoms)
# with sns.axes_style("white"):
fig, ax = plt.subplots(figsize=(2, 3.4))
# sns.palplot(sns.diverging_palette(240, 10, n=9))
ax = sns.heatmap(atoms, linewidths=.5, ax=ax, annot_kws={"size": 18}, cbar=False,
xticklabels=False, yticklabels=False, square=True, cmap="vlag", vmin=-1, vmax=1, linecolor='black')
# ,cmap=sns.diverging_palette(240, 10, n=9)) #"YlGnBu" , square=True
plt.show()
fig.savefig('atom.pdf')
fig.savefig('atom.png')
for i, x in enumerate(bond):
fig, ax = plt.subplots(figsize=(5, 5))
# sns.palplot(sns.diverging_palette(240, 10, n=9))
ax = sns.heatmap(x, linewidths=.5, ax=ax, annot_kws={"size": 18}, cbar=False,
xticklabels=False, yticklabels=False, square=True, cmap="vlag", vmin=-1, vmax=1, linecolor='black')
# ,cmap=sns.diverging_palette(240, 10, n=9)) #"YlGnBu" , square=True
plt.show()
fig.savefig('bond{}.pdf'.format(i))
fig.savefig('bond{}.png'.format(i))
if __name__ == '__main__':
# plot_mol()
# plot_mol_constraint_opt()
# plot_mol_matrix()
# plot_top_qed_mol()
# exit(-1)
start = time.time()
print("Start at Time: {}".format(time.ctime()))
parser = argparse.ArgumentParser()
parser.add_argument("--data_dir", type=str, default='data')
parser.add_argument('--data_name', type=str, choices=['qm9', 'zinc250k'], required=True,
help='dataset name')
parser.add_argument("--snapshot_path", "-snapshot", type=str, default='model_snapshot_epoch_200')
parser.add_argument("--hyperparams_path", type=str, default='moflow-params.json')
parser.add_argument("--property_model_path", type=str, default=None)
parser.add_argument("--save_path", type=str, required=True)
parser.add_argument("--batch_size", type=int, default=256)
parser.add_argument('-l', '--learning_rate', type=float, default=0.001, help='Base learning rate')
parser.add_argument('-e', '--lr_decay', type=float, default=0.999995,
help='Learning rate decay, applied every step of the optimization')
parser.add_argument('-w', '--weight_decay', type=float, default=1e-5,
help='L2 norm for the parameters')
parser.add_argument('--hidden', type=str, default="",
help='Hidden dimension list for output regression')
parser.add_argument('-x', '--max_epochs', type=int, default=5, help='How many epochs to run in total?')
parser.add_argument('-g', '--gpu', type=int, default=0, help='GPU Id to use')
parser.add_argument("--delta", type=float, default=0.01)
parser.add_argument("--img_format", type=str, default='svg')
parser.add_argument('--multi_property', action='store_true', default=False, help='To run optimization/constrained optimization with multiple properties')
parser.add_argument('--property_name', type=str, default='plogp', choices=['qed', 'plogp', 'drd2', 'gsk3b', 'jnk3', 'qed_plogp'])
parser.add_argument('--additive_transformations', type=strtobool, default=False,
help='apply only additive coupling layers')
parser.add_argument('--temperature', type=float, default=1.0,
help='temperature of the gaussian distributions')
parser.add_argument('--topk', type=int, default=800, help='Top k smiles as seeds')
parser.add_argument('--path_range', type=int, default=100, help='Range of manipulation')
parser.add_argument('--debug', type=strtobool, default='true', help='To run optimization with more information')
parser.add_argument("--sim_cutoff", type=float, default=0.00)
parser.add_argument('--topscore', action='store_true', default=False, help='To find top score')
parser.add_argument('--consopt', action='store_true', default=False, help='To do constrained optimization')
args = parser.parse_args()
# Device configuration
device = -1
if args.gpu >= 0:
# device = args.gpu
device = torch.device('cuda:' + str(args.gpu) if torch.cuda.is_available() else 'cpu')
else:
device = torch.device('cpu')
property_name = args.property_name.lower()
if args.data_name == 'qm9':
qm9_model = 'models/results/qm9_64gnn_128-64lin_1-1mask_0d6noise_convlu1'
model_dir = os.path.join(os.getcwd(), qm9_model)
elif args.data_name == 'zinc250k':
zinc_model = 'models/results/zinc250k_512t2cnn_256gnn_512-64lin_10flow_19fold_convlu2_38af-1-1mask'
model_dir = os.path.join(os.getcwd(), zinc_model)
snapshot_path = os.path.join(model_dir, args.snapshot_path)
hyperparams_path = os.path.join(model_dir, args.hyperparams_path)
model_params = Hyperparameters(path=hyperparams_path)
model = load_model(snapshot_path, model_params, debug=True) # Load moflow model
if args.hidden in ('', ','):
hidden = []
else:
hidden = [int(d) for d in args.hidden.strip(',').split(',')]
print('Hidden dim for output regression: ', hidden)
if args.data_name == 'qm9':
atomic_num_list = [6, 7, 8, 9, 0]
transform_fn = transform_qm9.transform_fn
valid_idx = transform_qm9.get_val_ids()
molecule_file = 'qm9_relgcn_kekulized_ggnp.npz'
elif args.data_name == 'zinc250k':
atomic_num_list = zinc250_atomic_num_list
transform_fn = transform_zinc250k.transform_fn_zinc250k
valid_idx = transform_zinc250k.get_val_ids()
molecule_file = 'zinc250k_relgcn_kekulized_ggnp.npz'
else:
raise ValueError("Wrong data_name{}".format(args.data_name))
# dataset = NumpyTupleDataset(os.path.join(args.data_dir, molecule_file), transform=transform_fn) # 133885
dataset = NumpyTupleDataset.load(os.path.join(args.data_dir, molecule_file), transform=transform_fn)
print('Load {} done, length: {}'.format(os.path.join(args.data_dir, molecule_file), len(dataset)))
assert len(valid_idx) > 0
train_idx = [t for t in range(len(dataset)) if t not in valid_idx] # 224568 = 249455 - 24887
n_train = len(train_idx) # 120803 zinc: 224568
train = torch.utils.data.Subset(dataset, train_idx) # 120803
test = torch.utils.data.Subset(dataset, valid_idx) # 13082 not used for generation
train_dataloader = torch.utils.data.DataLoader(train, batch_size=args.batch_size)
# print("loading hyperparamaters from {}".format(hyperparams_path))
prop_list, prop_to_idx = load_property_csv(args.data_name, normalize=False)
train_prop = [prop_list[i] for i in train_idx]
test_prop = [prop_list[i] for i in valid_idx]
print('Prepare data done! Time {:.2f} seconds'.format(time.time() - start))
model.to(device)
model.eval()
if args.topscore:
print('Finding top score:')
if not args.multi_property:
find_top_score_smiles(model, device, args.data_name, property_name, prop_to_idx, train_prop, args.topk, atomic_num_list, args.save_path)
else:
find_top_score_smiles_multi_prop(model, device, args.data_name, property_name, prop_to_idx, train_prop, args.topk, atomic_num_list, args.save_path)
if args.consopt:
print('Constrained optimization:')
if not args.multi_property:
constrain_optimization_smiles(model, device, args.data_name, property_name, prop_to_idx, train_prop, args.topk,
atomic_num_list, args.save_path, path_range=args.path_range, sim_cutoff=args.sim_cutoff)
else:
constrain_optimization_smiles_multi_prop(model, device, args.data_name, property_name, prop_to_idx, train_prop, args.topk,
atomic_num_list, args.save_path, path_range=args.path_range, sim_cutoff=args.sim_cutoff)
print('Total Time {:.2f} seconds'.format(time.time() - start))