File size: 9,044 Bytes
d05f89f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
# -*- coding: utf-8 -*-
# The B3clf library computes the blood-brain barrier (BBB) permeability
# of organic molecules with resampling strategies.
#
# Copyright (C) 2021 The Ayers Lab
#
# This file is part of B3clf.
#
# B3clf is free software; you can redistribute it and/or
# modify it under the terms of the GNU General Public License
# as published by the Free Software Foundation; either version 3
# of the License, or (at your option) any later version.
#
# B3clf is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, see <http://www.gnu.org/licenses/>
#
# --

import pandas as pd
from rdkit import Chem
from rdkit.Chem import AllChem

"""Convert SMILES to 3D and/or minimize the geometry from SDF with force field."""


def geometry_optimize(input_fname,
                      output_sdf,
                      steps_opt=10000,
                      # convergence=1.e-7,
                      tool="rdkit",
                      # optimization="cg",
                      force_field="MMFF94s",
                      smi_col=None,
                      sep="\s+|t+"):
    """Generate 3D coordinates and run geometry optimization with force field."""

    # optimize the 3d coordinates
    # use RDKit to minimize the geometry
    if tool.lower() == "rdkit":
        minimize_with_rdkit(input_molfname=input_fname,
                            sdf_out=output_sdf,
                            maxIters=steps_opt,
                            force_field=force_field,
                            smi_col=smi_col,
                            sep=sep)
    # use openbabel to minimize the geometry
    elif tool == "openbabel":
        # minimize_with_openbabel(input_molfname=input_fname,
        #                         sdf_out=output_sdf,
        #                         steps=steps_opt,
        #                         optimization=optimization,
        #                         convergence=convergence,
        #                         force_field=force_field,
        #                         smi_col=smi_col)
        raise ValueError("OpenBabel is not supported yet.")
    else:
        raise ValueError("{} not implemented yet.".format(tool))


def minimize_with_rdkit(input_molfname,
                        sdf_out,
                        smi_col=None,
                        mol_name_col=None,
                        maxIters=400,
                        force_field="MMFF94s",
                        sep="\s+"):
    """Add hydrogen for 3D coordinates and minimize the geometry with RdKit."""
    # load molecules
    if input_molfname.lower().endswith(".smi") or input_molfname.lower().endswith(".csv"):
        # todo: support .txt files
        # todo: add support of more flexible separators
        # todo: fix problem when mol_name is empty
        df_mol = pd.read_csv(input_molfname, sep=sep, engine="python", header=None)
        if df_mol.shape[1] == 1:
            # Case for only SMILES column
            smile_list = df_mol.iloc[:, -1].to_list()
            mol_name_list = df_mol.iloc[:, -1].to_list()
        else:
            # Case for SMILES and MOL name columns
            if smi_col is None:
                smile_list = df_mol.iloc[:, 0].to_list()
            else:
                smile_list = df_mol[smi_col].to_list()

            if mol_name_col is None:
            # todo: use name if column name is valid
                mol_name_list = df_mol.iloc[:, -1].to_list()
            else:
                mol_name_list = df_mol[mol_name_col].to_list()

        mols = []
        for idx, smi in enumerate(smile_list):
            mol = Chem.MolFromSmiles(smi)
            # This will overwrite
            if mol is not None:
                mol.SetProp("_Name", mol_name_list[idx])
                mols.append(mol)

    elif input_molfname.lower().endswith(".sdf"):
        suppl = Chem.SDMolSupplier(input_molfname,
                                   sanitize=True,
                                   removeHs=False,
                                   strictParsing=True)
        mols = [mol for mol in suppl]
        for idx, mol in enumerate(mols):
            if (mol.GetProp("_Name") == "") or (mol.GetProp("_Name") is None):
                smi = Chem.MolToSmiles(mol)
                mol.SetProp("_Name", smi)
                mols[idx] = mol

    writer = Chem.SDWriter(sdf_out)
    for idx, mol in enumerate(mols):
        mol = Chem.AddHs(mol)
        if force_field == "MMFF94s":
            # use MMFF~ force field if possible

            # taken from
            # https://open-babel.readthedocs.io/en/latest/Forcefields/mmff94.html
            # Some experiments and most theoretical calculations show significant pyramidal
            # “puckering” at nitrogens in isolated structures. The MMFF94s (static) variant has
            # slightly different  out-of-plane bending and dihedral torsion parameters to planarize
            # certain types of delocalized trigonal N atoms, such as aromatic aniline. This provides
            # a better match to the time-average molecular geometry in solution or crystal
            # structures.
            #
            # If you are comparing force-field optimized molecules to crystal structure geometries,
            # we recommend using the MMFF94s variant for this reason. All other parameters are
            # identical. However, if you are performing “docking” simulations, consideration of
            # active solution conformations, or other types of computational studies, we recommend
            # using the MMFF94 variant, since one form or another of the N geometry will
            # predominate.

            AllChem.EmbedMolecule(mol, randomSeed=999)
            # the following code will raise some errors
            mini_tag = AllChem.MMFFOptimizeMolecule(mol, force_field, maxIters=maxIters)
            # 0 optimize converged
            # -1 can not set up force field
            # 1 more iterations required
            if mini_tag == 0:
                writer.write(mol)
            else:
                if mini_tag == 1:
                    AllChem.MMFFOptimizeMolecule(mol, force_field, maxIters=maxIters * 2)
                elif mini_tag == -1:
                    AllChem.UFFOptimizeMolecule(mol, maxIters=400)
                writer.write(mol)

        elif force_field == "uff":
            # use uff force field if possible
            AllChem.EmbedMolecule(mol, randomSeed=999)
            # the following code will raise some errors
            mini_tag = AllChem.UFFOptimizeMolecule(mol, maxIters=maxIters)
            # 0 optimize converged
            # -1 can not set up force field
            # 1 more iterations required
            if mini_tag == 0:
                writer.write(mol)
            else:
                if mini_tag == 1:
                    AllChem.UFFOptimizeMolecule(mol, maxIters=maxIters * 2)
                elif mini_tag == -1:
                    AllChem.MMFFOptimizeMolecule(mol, "MMFF94s", maxIters=maxIters)
                writer.write(mol)

        else:
            raise NotImplementedError("This method is not implemented yet.")

    writer.close()

# todo: now the implementation is not supporting adding molecule name (such as SMILES strings)
# def minimize_with_openbabel(input_molfname,
#                             sdf_out,
#                             steps=10000,
#                             convergence=1.e-7,
#                             optimization="cg",
#                             force_field="GAFF",
#                             smi_col=None):
#     """Minimize the geometries with openbabel.
#
#     Parameters
#     ----------
#     input_molfname : str
#         Input molecule fie name.
#     sdf_out : str
#         Output molecule file name.
#     steps : int, optional
#         Specify the maximum number of steps. default=2500.
#     optimization : str, optional
#         Use conjugate gradients ("cg") or steepest descent ("sd") algorithm for optimization.
#         Default="cg".
#     convergence : float, optional
#         convergence threshold. Default=1.e-7.
#     force_field : str, optional
#         ForceField name including Generalized Amber Force Field (gaff), Ghemical Force Field
#         (ghemical), MMFF94 Force Field (mmff94) and Universal Force Field (uff). Default="gaff".
#     """
#
#     # https://open-babel.readthedocs.io/en/latest/Command-line_tools/babel.html#forcefield-energy-and-minimization
#     subprocess.Popen(["obabel", input_molfname, "-h", "-O", sdf_out,
#                       "--gen3d", "--minimize",
#                       "--n", str(steps), "--sd", optimization, "--crit",
#                       str(convergence), "--ff", force_field])
#     print("Geometry optimization with OpenBabel is done.")