-
Notifications
You must be signed in to change notification settings - Fork 171
/
graph_ga_run.py
271 lines (204 loc) · 9.15 KB
/
graph_ga_run.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
from __future__ import print_function
import argparse
import heapq
import json
import os
import random
from time import time
from typing import List, Optional
from tqdm import tqdm
import pickle
import joblib
import numpy as np
# from guacamol.assess_goal_directed_generation import assess_goal_directed_generation
# from guacamol.goal_directed_generator import GoalDirectedGenerator
# from guacamol.scoring_function import ScoringFunction
# from guacamol.utils.helpers import setup_default_logger
# from joblib import delayed
from rdkit import Chem
from rdkit.Chem.rdchem import Mol
from graph_ga import crossover as co, mutate as mu
result_folder = '/project/molecular_data/graphnn/pyscreener/graph_ga/results.3/'
from tdc import Oracle
drd3_oracle = Oracle(name = 'Docking_Score', software='vina',
pyscreener_path = '/project/molecular_data/graphnn/pyscreener',
receptors=['/project/molecular_data/graphnn/pyscreener/testing_inputs/DRD3.pdb'],
center=(9, 22.5, 26), size=(15, 15, 15),
buffer=10, path='/project/molecular_data/graphnn/pyscreener/my_test/', num_worker=1, ncpu=10)
qed_oracle = Oracle(name='qed')
# def drd3_docking_oracle(smiles):
# return max(-drd3_oracle(smiles),0)
def drd3_docking_oracle(smiles):
qed = qed_oracle(smiles)
if qed < 0.2:
return 0.0
return max(-drd3_oracle(smiles),0)
# def drd3_docking_oracle(smiles):
# return random.random()
def canonicalize(smiles: str, include_stereocenters=True):
"""
Canonicalize the SMILES strings with RDKit.
The algorithm is detailed under https://pubs.acs.org/doi/full/10.1021/acs.jcim.5b00543
Args:
smiles: SMILES string to canonicalize
include_stereocenters: whether to keep the stereochemical information in the canonical SMILES string
Returns:
Canonicalized SMILES string, None if the molecule is invalid.
"""
mol = Chem.MolFromSmiles(smiles)
if mol is not None:
return Chem.MolToSmiles(mol, isomericSmiles=include_stereocenters)
else:
return None
def read_in_num(filename):
with open(filename, 'r') as fin:
line = fin.readline()
return int(line)
def make_mating_pool(population_mol: List[Mol], population_scores, offspring_size: int):
"""
Given a population of RDKit Mol and their scores, sample a list of the same size
with replacement using the population_scores as weights
Args:
population_mol: list of RDKit Mol
population_scores: list of un-normalised scores given by ScoringFunction
offspring_size: number of molecules to return
Returns: a list of RDKit Mol (probably not unique)
"""
# scores -> probs
mol_list, score_list = [], []
for mol, score in zip(population_mol, population_scores):
# if 0<= score <=1:
mol_list.append(mol)
score_list.append(score)
# print("score_list", score_list)
population_mol, population_scores = mol_list, score_list
sum_scores = sum(population_scores)
if sum_scores==0:
population_probs = [1.0/len(population_scores) for i in range(len(population_scores))]
else:
population_probs = [p / sum_scores for p in population_scores]
# print(population_probs, 'population_probs')
mating_pool = np.random.choice(population_mol, p=population_probs, size=offspring_size, replace=True)
return mating_pool
def reproduce(mating_pool, mutation_rate):
"""
Args:
mating_pool: list of RDKit Mol
mutation_rate: rate of mutation
Returns:
"""
parent_a = random.choice(mating_pool)
parent_b = random.choice(mating_pool)
new_child = co.crossover(parent_a, parent_b)
if new_child is not None:
new_child = mu.mutate(new_child, mutation_rate)
return new_child
def score_mol(mol, score_fn):
return score_fn(Chem.MolToSmiles(mol))
def sanitize(population_mol):
new_population = []
smile_set = set()
for mol in population_mol:
if mol is not None:
try:
smile = Chem.MolToSmiles(mol)
if smile is not None and smile not in smile_set:
smile_set.add(smile)
new_population.append(mol)
except ValueError:
print('bad smiles')
return new_population
class GB_GA_Generator:
def __init__(self, smi_file, population_size, offspring_size, generations, mutation_rate):
self.smi_file = smi_file
self.all_smiles = self.load_smiles_from_file(self.smi_file)
self.population_size = population_size
self.offspring_size = offspring_size
self.generations = generations
self.mutation_rate = mutation_rate
self.smiles2score = {} ### save the result up to now.
self.oracle_num = 0
def load_smiles_from_file(self, smi_file):
with open(smi_file) as f:
lines = f.readlines()
return [line.strip() for line in lines]
def generate_optimized_molecules(self):
from random import shuffle
shuffle(self.all_smiles)
starting_population = self.all_smiles[:self.population_size]
starting_population = [canonicalize(smiles) for smiles in starting_population]
print("---------- before init, oracle_num", self.oracle_num, 'population_size', self.population_size)
# select initial population
# population_smiles = heapq.nlargest(self.population_size, starting_population, key=oracle)
population_smiles = starting_population
population_mol = [Chem.MolFromSmiles(s) for s in population_smiles]
# population_scores = self.pool(delayed(score_mol)(m, scoring_function.score) for m in population_mol)
population_scores = []
for smiles in population_smiles:
smiles = canonicalize(smiles)
if smiles in self.smiles2score:
score = self.smiles2score[smiles]
else:
try:
score = drd3_docking_oracle(smiles)
except:
score = 0
self.oracle_num += 1
print('docking number', self.oracle_num)
if self.oracle_num % 100 == 0:
pickle.dump(self.smiles2score, open(result_folder + str(self.oracle_num) + '.pkl', 'wb'))
self.smiles2score[smiles] = score
population_scores.append(score)
for generation in tqdm(range(self.generations)): ### 1k
# new_population
mating_pool = make_mating_pool(population_mol, population_scores, self.offspring_size)
# offspring_mol = self.pool(delayed(reproduce)(mating_pool, self.mutation_rate) for _ in range(self.population_size))
offspring_mol = []
for _ in range(self.population_size):
offspring = reproduce(mating_pool, self.mutation_rate)
offspring_mol.append(offspring)
# add new_population
population_mol += offspring_mol
population_mol = sanitize(population_mol)
old_scores = population_scores
# population_scores = self.pool(delayed(score_mol)(m, scoring_function.score) for m in population_mol)
population_scores = []
for m in population_mol:
smiles = Chem.MolToSmiles(m)
smiles = canonicalize(smiles)
if smiles in self.smiles2score:
score = self.smiles2score[smiles]
else:
try:
score = drd3_docking_oracle(smiles)
except:
score = 0
self.oracle_num += 1
print('docking number', self.oracle_num)
if self.oracle_num % 100 == 0:
pickle.dump(self.smiles2score, open(result_folder + str(self.oracle_num) + '.pkl', 'wb'))
self.smiles2score[smiles] = score
population_scores.append(drd3_docking_oracle(Chem.MolToSmiles(m)))
population_tuples = list(zip(population_scores, population_mol))
population_tuples = sorted(population_tuples, key=lambda x: x[0], reverse=True)[:self.population_size]
population_mol = [t[1] for t in population_tuples]
population_scores = [t[0] for t in population_tuples]
return [Chem.MolToSmiles(m) for m in population_mol][:]
ga = GB_GA_Generator(smi_file = "data/guacamol_v1_all.smiles",
population_size = 100,
offspring_size = 200,
generations = 100000,
mutation_rate = 0.01, )
ga.generate_optimized_molecules()
num2score = {}
for file in os.listdir(result_folder):
num = int(file.split('.')[0])
smiles2score = pickle.load(open(result_folder + file, 'rb'))
smiles_score_lst = [(smiles,score) for smiles, score in smiles2score.items()]
smiles_score_lst.sort(key=lambda x:x[1], reverse=True)
score_lst = [i[1] for i in smiles_score_lst]
mean_score = np.mean(score_lst[:10])
num2score[num] = mean_score
num_score_lst = [(num,score) for num,score in num2score.items()]
num_score_lst.sort(key = lambda x:x[0])
print(num_score_lst)