ms-adtools 0.1.0__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- ms_adtools-0.1.0/.github/workflows/publish.yml +31 -0
- ms_adtools-0.1.0/.gitignore +220 -0
- ms_adtools-0.1.0/LICENSE +25 -0
- ms_adtools-0.1.0/PKG-INFO +122 -0
- ms_adtools-0.1.0/README.md +95 -0
- ms_adtools-0.1.0/benchmarks/REPORT.md +72 -0
- ms_adtools-0.1.0/benchmarks/bench.py +128 -0
- ms_adtools-0.1.0/pyproject.toml +46 -0
- ms_adtools-0.1.0/setup.cfg +4 -0
- ms_adtools-0.1.0/src/ms_adtools/__init__.py +15 -0
- ms_adtools-0.1.0/src/ms_adtools/ccd.py +221 -0
- ms_adtools-0.1.0/src/ms_adtools/cli.py +129 -0
- ms_adtools-0.1.0/src/ms_adtools/ligand.py +398 -0
- ms_adtools-0.1.0/src/ms_adtools/poses.py +168 -0
- ms_adtools-0.1.0/src/ms_adtools/reactive.py +118 -0
- ms_adtools-0.1.0/src/ms_adtools/receptor.py +569 -0
- ms_adtools-0.1.0/src/ms_adtools/zinc/AD4Zn.dat +121 -0
- ms_adtools-0.1.0/src/ms_adtools/zinc/GPL-2.0 +339 -0
- ms_adtools-0.1.0/src/ms_adtools/zinc/LICENSE +202 -0
- ms_adtools-0.1.0/src/ms_adtools/zinc/NOTICE +15 -0
- ms_adtools-0.1.0/src/ms_adtools/zinc/__init__.py +21 -0
- ms_adtools-0.1.0/src/ms_adtools/zinc/zinc_pseudo.py +565 -0
- ms_adtools-0.1.0/src/ms_adtools.egg-info/PKG-INFO +122 -0
- ms_adtools-0.1.0/src/ms_adtools.egg-info/SOURCES.txt +37 -0
- ms_adtools-0.1.0/src/ms_adtools.egg-info/dependency_links.txt +1 -0
- ms_adtools-0.1.0/src/ms_adtools.egg-info/entry_points.txt +2 -0
- ms_adtools-0.1.0/src/ms_adtools.egg-info/requires.txt +6 -0
- ms_adtools-0.1.0/src/ms_adtools.egg-info/scm_file_list.json +33 -0
- ms_adtools-0.1.0/src/ms_adtools.egg-info/scm_version.json +8 -0
- ms_adtools-0.1.0/src/ms_adtools.egg-info/top_level.txt +1 -0
- ms_adtools-0.1.0/test/data/1iep.pdbqt +2703 -0
- ms_adtools-0.1.0/test/data/1iep_flex_vina_out.pdbqt +348 -0
- ms_adtools-0.1.0/test/data/1iep_receptorH.pdb +4414 -0
- ms_adtools-0.1.0/test/data/3pg.pdb +12 -0
- ms_adtools-0.1.0/test/data/3pg_chem_comp.cif +118 -0
- ms_adtools-0.1.0/test/data/altloc.pdb +44 -0
- ms_adtools-0.1.0/test/data/ccd/3PG.cif +122 -0
- ms_adtools-0.1.0/test/data/zn_site.pdbqt +51 -0
- ms_adtools-0.1.0/test/test_prepare.py +282 -0
|
@@ -0,0 +1,31 @@
|
|
|
1
|
+
# Publish to PyPI on a GitHub release.
|
|
2
|
+
#
|
|
3
|
+
# No secrets: the job authenticates to PyPI over OIDC (Trusted Publishing).
|
|
4
|
+
# It has to be enabled on the PyPI side *before* the first upload -- for a
|
|
5
|
+
# project that does not exist yet, that is a "pending publisher" under
|
|
6
|
+
# Your projects > Publishing, and it must name this repo, `publish.yml`
|
|
7
|
+
# and the `pypi` environment below.
|
|
8
|
+
name: publish
|
|
9
|
+
|
|
10
|
+
on:
|
|
11
|
+
release:
|
|
12
|
+
types: [published]
|
|
13
|
+
|
|
14
|
+
jobs:
|
|
15
|
+
pypi:
|
|
16
|
+
runs-on: ubuntu-latest
|
|
17
|
+
environment: pypi
|
|
18
|
+
permissions:
|
|
19
|
+
# A permissions block sets every scope it does not name to none, so
|
|
20
|
+
# checkout needs contents spelled out or it cannot read a private repo.
|
|
21
|
+
contents: read
|
|
22
|
+
id-token: write # mint the OIDC token PyPI checks; no secret involved
|
|
23
|
+
steps:
|
|
24
|
+
- uses: actions/checkout@v7
|
|
25
|
+
with:
|
|
26
|
+
fetch-depth: 0 # setuptools-scm reads the version off the tag
|
|
27
|
+
- uses: actions/setup-python@v7
|
|
28
|
+
with:
|
|
29
|
+
python-version: "3.12"
|
|
30
|
+
- run: pipx run build
|
|
31
|
+
- uses: pypa/gh-action-pypi-publish@release/v1
|
|
@@ -0,0 +1,220 @@
|
|
|
1
|
+
# Byte-compiled / optimized / DLL files
|
|
2
|
+
__pycache__/
|
|
3
|
+
*.py[codz]
|
|
4
|
+
*$py.class
|
|
5
|
+
|
|
6
|
+
# C extensions
|
|
7
|
+
*.so
|
|
8
|
+
|
|
9
|
+
# Distribution / packaging
|
|
10
|
+
.Python
|
|
11
|
+
build/
|
|
12
|
+
develop-eggs/
|
|
13
|
+
dist/
|
|
14
|
+
downloads/
|
|
15
|
+
eggs/
|
|
16
|
+
.eggs/
|
|
17
|
+
lib/
|
|
18
|
+
lib64/
|
|
19
|
+
parts/
|
|
20
|
+
sdist/
|
|
21
|
+
var/
|
|
22
|
+
wheels/
|
|
23
|
+
share/python-wheels/
|
|
24
|
+
*.egg-info/
|
|
25
|
+
.installed.cfg
|
|
26
|
+
*.egg
|
|
27
|
+
MANIFEST
|
|
28
|
+
|
|
29
|
+
# PyInstaller
|
|
30
|
+
# Usually these files are written by a python script from a template
|
|
31
|
+
# before PyInstaller builds the exe, so as to inject date/other infos into it.
|
|
32
|
+
*.manifest
|
|
33
|
+
*.spec
|
|
34
|
+
|
|
35
|
+
# Installer logs
|
|
36
|
+
pip-log.txt
|
|
37
|
+
pip-delete-this-directory.txt
|
|
38
|
+
|
|
39
|
+
# Unit test / coverage reports
|
|
40
|
+
htmlcov/
|
|
41
|
+
.tox/
|
|
42
|
+
.nox/
|
|
43
|
+
.coverage
|
|
44
|
+
.coverage.*
|
|
45
|
+
.cache
|
|
46
|
+
nosetests.xml
|
|
47
|
+
coverage.xml
|
|
48
|
+
*.cover
|
|
49
|
+
*.py.cover
|
|
50
|
+
*.lcov
|
|
51
|
+
.hypothesis/
|
|
52
|
+
.pytest_cache/
|
|
53
|
+
cover/
|
|
54
|
+
|
|
55
|
+
# Translations
|
|
56
|
+
*.mo
|
|
57
|
+
*.pot
|
|
58
|
+
|
|
59
|
+
# Django stuff:
|
|
60
|
+
*.log
|
|
61
|
+
local_settings.py
|
|
62
|
+
db.sqlite3
|
|
63
|
+
db.sqlite3-journal
|
|
64
|
+
|
|
65
|
+
# Flask stuff:
|
|
66
|
+
instance/
|
|
67
|
+
.webassets-cache
|
|
68
|
+
|
|
69
|
+
# Scrapy stuff:
|
|
70
|
+
.scrapy
|
|
71
|
+
|
|
72
|
+
# Sphinx documentation
|
|
73
|
+
docs/_build/
|
|
74
|
+
|
|
75
|
+
# PyBuilder
|
|
76
|
+
.pybuilder/
|
|
77
|
+
target/
|
|
78
|
+
|
|
79
|
+
# Jupyter Notebook
|
|
80
|
+
.ipynb_checkpoints
|
|
81
|
+
|
|
82
|
+
# IPython
|
|
83
|
+
profile_default/
|
|
84
|
+
ipython_config.py
|
|
85
|
+
|
|
86
|
+
# pyenv
|
|
87
|
+
# For a library or package, you might want to ignore these files since the code is
|
|
88
|
+
# intended to run in multiple environments; otherwise, check them in:
|
|
89
|
+
# .python-version
|
|
90
|
+
|
|
91
|
+
# pipenv
|
|
92
|
+
# According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
|
|
93
|
+
# However, in case of collaboration, if having platform-specific dependencies or dependencies
|
|
94
|
+
# having no cross-platform support, pipenv may install dependencies that don't work, or not
|
|
95
|
+
# install all needed dependencies.
|
|
96
|
+
# Pipfile.lock
|
|
97
|
+
|
|
98
|
+
# uv
|
|
99
|
+
# Similar to Pipfile.lock, it is generally recommended to include uv.lock in version control.
|
|
100
|
+
# This is especially recommended for binary packages to ensure reproducibility, and is more
|
|
101
|
+
# commonly ignored for libraries.
|
|
102
|
+
# uv.lock
|
|
103
|
+
|
|
104
|
+
# poetry
|
|
105
|
+
# Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
|
|
106
|
+
# This is especially recommended for binary packages to ensure reproducibility, and is more
|
|
107
|
+
# commonly ignored for libraries.
|
|
108
|
+
# https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
|
|
109
|
+
# poetry.lock
|
|
110
|
+
# poetry.toml
|
|
111
|
+
|
|
112
|
+
# pdm
|
|
113
|
+
# Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
|
|
114
|
+
# pdm recommends including project-wide configuration in pdm.toml, but excluding .pdm-python.
|
|
115
|
+
# https://pdm-project.org/en/latest/usage/project/#working-with-version-control
|
|
116
|
+
# pdm.lock
|
|
117
|
+
# pdm.toml
|
|
118
|
+
.pdm-python
|
|
119
|
+
.pdm-build/
|
|
120
|
+
|
|
121
|
+
# pixi
|
|
122
|
+
# Similar to Pipfile.lock, it is generally recommended to include pixi.lock in version control.
|
|
123
|
+
# pixi.lock
|
|
124
|
+
# Pixi creates a virtual environment in the .pixi directory, just like venv module creates one
|
|
125
|
+
# in the .venv directory. It is recommended not to include this directory in version control.
|
|
126
|
+
.pixi/*
|
|
127
|
+
!.pixi/config.toml
|
|
128
|
+
|
|
129
|
+
# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
|
|
130
|
+
__pypackages__/
|
|
131
|
+
|
|
132
|
+
# Celery stuff
|
|
133
|
+
celerybeat-schedule*
|
|
134
|
+
celerybeat.pid
|
|
135
|
+
|
|
136
|
+
# Redis
|
|
137
|
+
*.rdb
|
|
138
|
+
*.aof
|
|
139
|
+
*.pid
|
|
140
|
+
|
|
141
|
+
# RabbitMQ
|
|
142
|
+
mnesia/
|
|
143
|
+
rabbitmq/
|
|
144
|
+
rabbitmq-data/
|
|
145
|
+
|
|
146
|
+
# ActiveMQ
|
|
147
|
+
activemq-data/
|
|
148
|
+
|
|
149
|
+
# SageMath parsed files
|
|
150
|
+
*.sage.py
|
|
151
|
+
|
|
152
|
+
# Environments
|
|
153
|
+
.env
|
|
154
|
+
.envrc
|
|
155
|
+
.venv
|
|
156
|
+
env/
|
|
157
|
+
venv/
|
|
158
|
+
ENV/
|
|
159
|
+
env.bak/
|
|
160
|
+
venv.bak/
|
|
161
|
+
|
|
162
|
+
# Spyder project settings
|
|
163
|
+
.spyderproject
|
|
164
|
+
.spyproject
|
|
165
|
+
|
|
166
|
+
# Rope project settings
|
|
167
|
+
.ropeproject
|
|
168
|
+
|
|
169
|
+
# mkdocs/Zensical documentation
|
|
170
|
+
/site
|
|
171
|
+
|
|
172
|
+
# mypy
|
|
173
|
+
.mypy_cache/
|
|
174
|
+
.dmypy.json
|
|
175
|
+
dmypy.json
|
|
176
|
+
|
|
177
|
+
# Pyre type checker
|
|
178
|
+
.pyre/
|
|
179
|
+
|
|
180
|
+
# pytype static type analyzer
|
|
181
|
+
.pytype/
|
|
182
|
+
|
|
183
|
+
# Cython debug symbols
|
|
184
|
+
cython_debug/
|
|
185
|
+
|
|
186
|
+
# PyCharm
|
|
187
|
+
# JetBrains specific template is maintained in a separate JetBrains.gitignore that can
|
|
188
|
+
# be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
|
|
189
|
+
# and can be added to the global gitignore or merged into this file. For a more nuclear
|
|
190
|
+
# option (not recommended) you can uncomment the following to ignore the entire idea folder.
|
|
191
|
+
.idea/
|
|
192
|
+
|
|
193
|
+
# Abstra
|
|
194
|
+
# Abstra is an AI-powered process automation framework.
|
|
195
|
+
# Ignore directories containing user credentials, local state, and settings.
|
|
196
|
+
# Learn more at https://abstra.io/docs
|
|
197
|
+
.abstra/
|
|
198
|
+
|
|
199
|
+
# Visual Studio Code
|
|
200
|
+
# Visual Studio Code specific template is maintained in a separate VisualStudioCode.gitignore that
|
|
201
|
+
# can be found at https://github.com/github/gitignore/blob/main/Global/VisualStudioCode.gitignore
|
|
202
|
+
# and can be added to the global gitignore or merged into this file. However, if you prefer, you
|
|
203
|
+
# could uncomment the following to ignore the entire vscode folder
|
|
204
|
+
# .vscode/
|
|
205
|
+
# Temporary file for partial code execution
|
|
206
|
+
tempCodeRunnerFile.py
|
|
207
|
+
|
|
208
|
+
# Ruff stuff:
|
|
209
|
+
.ruff_cache/
|
|
210
|
+
|
|
211
|
+
# PyPI configuration file
|
|
212
|
+
.pypirc
|
|
213
|
+
|
|
214
|
+
# Marimo
|
|
215
|
+
marimo/_static/
|
|
216
|
+
marimo/_lsp/
|
|
217
|
+
__marimo__/
|
|
218
|
+
|
|
219
|
+
# Streamlit
|
|
220
|
+
.streamlit/secrets.toml
|
ms_adtools-0.1.0/LICENSE
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
1
|
+
MIT License
|
|
2
|
+
|
|
3
|
+
Copyright (c) 2026 Mario S. Valdés-Tresanco and Mario E. Valdés-Tresanco
|
|
4
|
+
|
|
5
|
+
Permission is hereby granted, free of charge, to any person obtaining a copy
|
|
6
|
+
of this software and associated documentation files (the "Software"), to deal
|
|
7
|
+
in the Software without restriction, including without limitation the rights
|
|
8
|
+
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
|
9
|
+
copies of the Software, and to permit persons to whom the Software is
|
|
10
|
+
furnished to do so, subject to the following conditions:
|
|
11
|
+
|
|
12
|
+
The above copyright notice and this permission notice shall be included in all
|
|
13
|
+
copies or substantial portions of the Software.
|
|
14
|
+
|
|
15
|
+
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
16
|
+
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
17
|
+
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
|
18
|
+
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
19
|
+
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
|
20
|
+
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
|
21
|
+
SOFTWARE.
|
|
22
|
+
|
|
23
|
+
The files in src/ms_adtools/zinc/ keep their own licenses: zinc_pseudo.py is
|
|
24
|
+
Apache-2.0 (see its LICENSE) and AD4Zn.dat is GPL-2.0-or-later data, shipped
|
|
25
|
+
unmodified alongside this package (see its NOTICE and GPL-2.0).
|
|
@@ -0,0 +1,122 @@
|
|
|
1
|
+
Metadata-Version: 2.4
|
|
2
|
+
Name: ms_adtools
|
|
3
|
+
Version: 0.1.0
|
|
4
|
+
Summary: AutoDock receptor, flexible-residue and ligand preparation with RDKit and gemmi.
|
|
5
|
+
Author: Mario S. Valdés-Tresanco, Mario E. Valdés-Tresanco
|
|
6
|
+
License-Expression: MIT
|
|
7
|
+
Project-URL: Homepage, https://github.com/MolSuite/ms_adtools
|
|
8
|
+
Project-URL: Issues, https://github.com/MolSuite/ms_adtools/issues
|
|
9
|
+
Keywords: autodock,vina,pdbqt,docking,molsuite
|
|
10
|
+
Classifier: Development Status :: 3 - Alpha
|
|
11
|
+
Classifier: Intended Audience :: Science/Research
|
|
12
|
+
Classifier: Programming Language :: Python :: 3 :: Only
|
|
13
|
+
Classifier: Programming Language :: Python :: 3.12
|
|
14
|
+
Classifier: Topic :: Scientific/Engineering :: Chemistry
|
|
15
|
+
Requires-Python: >=3.12
|
|
16
|
+
Description-Content-Type: text/markdown
|
|
17
|
+
License-File: LICENSE
|
|
18
|
+
License-File: src/ms_adtools/zinc/LICENSE
|
|
19
|
+
License-File: src/ms_adtools/zinc/GPL-2.0
|
|
20
|
+
License-File: src/ms_adtools/zinc/NOTICE
|
|
21
|
+
Requires-Dist: numpy>=2.0
|
|
22
|
+
Requires-Dist: gemmi>=0.7
|
|
23
|
+
Requires-Dist: rdkit>=2024.3
|
|
24
|
+
Provides-Extra: dev
|
|
25
|
+
Requires-Dist: pytest>=7.0; extra == "dev"
|
|
26
|
+
Dynamic: license-file
|
|
27
|
+
|
|
28
|
+
# ms_adtools
|
|
29
|
+
|
|
30
|
+
AutoDock preparation on RDKit and gemmi: ligand PDBQT, receptor PDBQT (PDB, mmCIF, PQR, PDBQT,
|
|
31
|
+
mol2), the rigid/flexible pair for flexible-residue docking as AutoDockTools writes it, and
|
|
32
|
+
docked poses (PDBQT, DLG) back to RDKit. Requires only `numpy`, `gemmi` and `rdkit`.
|
|
33
|
+
|
|
34
|
+
```bash
|
|
35
|
+
pip install ms_adtools
|
|
36
|
+
```
|
|
37
|
+
|
|
38
|
+
## Why
|
|
39
|
+
|
|
40
|
+
Performance. A virtual screen prepares molecules in many parallel workers, so what each worker
|
|
41
|
+
loads matters as much as how fast it runs. meeko, the reference preparation tool for AutoDock,
|
|
42
|
+
brings heavy dependencies (scipy, pandas, prody): importing it alone takes about 240 MB, and
|
|
43
|
+
preparing a receptor peaks above 350 MB. ms_adtools follows its ligand conventions (typing,
|
|
44
|
+
torsion tree, `REMARK SMILES` / `IDX` / `H PARENT`), so its ligand PDBQT reads the same, with numpy, gemmi
|
|
45
|
+
and RDKit as the only requirements: a worker that prepares a receptor peaks at about 80 MB.
|
|
46
|
+
|
|
47
|
+
## Command line
|
|
48
|
+
|
|
49
|
+
```bash
|
|
50
|
+
ms_adtools ligand lig.sdf -o lig.pdbqt # SDF, MOL, mol2, or .smi (embedded in 3D)
|
|
51
|
+
ms_adtools ligand library.smi -o library/ # several molecules: one PDBQT each
|
|
52
|
+
ms_adtools receptor 1iep.pdb -o 1iep.pdbqt # --repairs hydrogens, --zinc, --ccd, --charges
|
|
53
|
+
ms_adtools flexreceptor 1iep.pdbqt -r "A:239,315" # writes 1iep_rigid.pdbqt + 1iep_flex.pdbqt
|
|
54
|
+
ms_adtools poses docked.pdbqt -o poses.sdf # or a .dlg; one record per pose, with its score
|
|
55
|
+
```
|
|
56
|
+
|
|
57
|
+
`ms_adtools <command> -h` lists every option. A molecule that fails is reported on stderr and the
|
|
58
|
+
rest of the library goes on (exit code 1 if any failed).
|
|
59
|
+
|
|
60
|
+
## Selecting residues
|
|
61
|
+
|
|
62
|
+
`-r` / `residues=` takes chains and residue numbers, comma- or space-separated:
|
|
63
|
+
|
|
64
|
+
| selection | means |
|
|
65
|
+
|---|---|
|
|
66
|
+
| `A:2-10,B:20` | residues 2 to 10 of chain A, 20 of chain B |
|
|
67
|
+
| `A:2,4,6` | three residues of chain A (a chain holds until the next one) |
|
|
68
|
+
| `A:ARG239,THR315` | with names, which must then match |
|
|
69
|
+
| `100` / `100A` | any chain; `100` takes its insertion codes, `100A` only that one |
|
|
70
|
+
|
|
71
|
+
Residues with nothing to rotate (Gly, Ala) and Pro stay rigid, with a warning.
|
|
72
|
+
|
|
73
|
+
## Python API
|
|
74
|
+
|
|
75
|
+
```python
|
|
76
|
+
from rdkit import Chem
|
|
77
|
+
from rdkit.Chem import AllChem
|
|
78
|
+
from ms_adtools import prepare_ligand, prepare_receptor, prepare_flexreceptor, pdbqt_to_rdkit
|
|
79
|
+
|
|
80
|
+
mol = Chem.AddHs(Chem.MolFromSmiles("CC(=O)Nc1ccc(O)cc1"))
|
|
81
|
+
AllChem.EmbedMolecule(mol)
|
|
82
|
+
pdbqt = prepare_ligand(mol) # str, with REMARK SMILES / IDX / H PARENT
|
|
83
|
+
receptor = prepare_receptor("1iep.pdb") # str
|
|
84
|
+
rigid, flex = prepare_flexreceptor("1iep.pdbqt", "A:239,315")
|
|
85
|
+
poses = pdbqt_to_rdkit(docked_text) # one conformer per MODEL, `score` on each
|
|
86
|
+
```
|
|
87
|
+
|
|
88
|
+
| function | returns |
|
|
89
|
+
|---|---|
|
|
90
|
+
| `prepare_ligand(mol, open_macrocycles=False, hydrate=False, flexible_amides=False, ...)` | ligand PDBQT text |
|
|
91
|
+
| `prepare_reactive_ligand(mol, smarts, idx=0, ...)` | ligand PDBQT with the reactive atom typed |
|
|
92
|
+
| `prepare_receptor(path, repairs="", charges=None, cleanup=..., zinc=False, ccd=False)` | receptor PDBQT text |
|
|
93
|
+
| `prepare_flexreceptor(path, residues, reactive=())` | `(rigid, flex)` PDBQT texts |
|
|
94
|
+
| `pdbqt_to_rdkit(text)` | RDKit molecule, one conformer per pose |
|
|
95
|
+
| `dlg_to_pdbqt(text, cluster_leads=True)` | PDBQT poses of an AutoDock4/AutoDock-GPU DLG |
|
|
96
|
+
| `flexres_models(text, rigid=None)` | docked flexible residues as a multi-model PDB |
|
|
97
|
+
| `add_zinc_pseudo_atoms(pdbqt)` | receptor PDBQT with AutoDock4Zn TZ atoms |
|
|
98
|
+
|
|
99
|
+
`ms_adtools.reactive.reactive_config` writes the config of reactive docking;
|
|
100
|
+
`ms_adtools.zinc.PARAMETERS` is the `AD4Zn.dat` that autogrid4 needs for AutoDock4Zn maps.
|
|
101
|
+
Each function's docstring has the details.
|
|
102
|
+
|
|
103
|
+
## Speed and memory
|
|
104
|
+
|
|
105
|
+
One CPU, Python 3.12, RDKit 2026.09; each row is a fresh process.
|
|
106
|
+
|
|
107
|
+
| | wall | peak RSS |
|
|
108
|
+
|---|---|---|
|
|
109
|
+
| import | 0.12 s | 71 MB |
|
|
110
|
+
| 200 ligands (Enamine, 3D) | 0.29 s (1.5 ms each) | 72 MB |
|
|
111
|
+
| receptor (4uwh, 4905 atoms) | 0.16 s | 83 MB |
|
|
112
|
+
|
|
113
|
+
On `1iep_receptorH.pdb` the receptor has AutoDockTools' atoms and types (charges within 0.02 on
|
|
114
|
+
average), and `prepare_flexreceptor` writes its rigid file byte for byte. `benchmarks/` has the
|
|
115
|
+
script and the full measurements.
|
|
116
|
+
|
|
117
|
+
## License
|
|
118
|
+
|
|
119
|
+
MIT (`LICENSE`). The PDBQT conventions are AutoDockTools' (MGLTools, M. Sanner, TSRI) and the Forli
|
|
120
|
+
lab's; cite them with AutoDock. `src/ms_adtools/zinc/` carries AutoDock Vina's
|
|
121
|
+
`zinc_pseudo.py` (Apache-2.0) and the unmodified `AD4Zn.dat` parameter file (GPL-2.0-or-later
|
|
122
|
+
data), each with its own license (see its `NOTICE`).
|
|
@@ -0,0 +1,95 @@
|
|
|
1
|
+
# ms_adtools
|
|
2
|
+
|
|
3
|
+
AutoDock preparation on RDKit and gemmi: ligand PDBQT, receptor PDBQT (PDB, mmCIF, PQR, PDBQT,
|
|
4
|
+
mol2), the rigid/flexible pair for flexible-residue docking as AutoDockTools writes it, and
|
|
5
|
+
docked poses (PDBQT, DLG) back to RDKit. Requires only `numpy`, `gemmi` and `rdkit`.
|
|
6
|
+
|
|
7
|
+
```bash
|
|
8
|
+
pip install ms_adtools
|
|
9
|
+
```
|
|
10
|
+
|
|
11
|
+
## Why
|
|
12
|
+
|
|
13
|
+
Performance. A virtual screen prepares molecules in many parallel workers, so what each worker
|
|
14
|
+
loads matters as much as how fast it runs. meeko, the reference preparation tool for AutoDock,
|
|
15
|
+
brings heavy dependencies (scipy, pandas, prody): importing it alone takes about 240 MB, and
|
|
16
|
+
preparing a receptor peaks above 350 MB. ms_adtools follows its ligand conventions (typing,
|
|
17
|
+
torsion tree, `REMARK SMILES` / `IDX` / `H PARENT`), so its ligand PDBQT reads the same, with numpy, gemmi
|
|
18
|
+
and RDKit as the only requirements: a worker that prepares a receptor peaks at about 80 MB.
|
|
19
|
+
|
|
20
|
+
## Command line
|
|
21
|
+
|
|
22
|
+
```bash
|
|
23
|
+
ms_adtools ligand lig.sdf -o lig.pdbqt # SDF, MOL, mol2, or .smi (embedded in 3D)
|
|
24
|
+
ms_adtools ligand library.smi -o library/ # several molecules: one PDBQT each
|
|
25
|
+
ms_adtools receptor 1iep.pdb -o 1iep.pdbqt # --repairs hydrogens, --zinc, --ccd, --charges
|
|
26
|
+
ms_adtools flexreceptor 1iep.pdbqt -r "A:239,315" # writes 1iep_rigid.pdbqt + 1iep_flex.pdbqt
|
|
27
|
+
ms_adtools poses docked.pdbqt -o poses.sdf # or a .dlg; one record per pose, with its score
|
|
28
|
+
```
|
|
29
|
+
|
|
30
|
+
`ms_adtools <command> -h` lists every option. A molecule that fails is reported on stderr and the
|
|
31
|
+
rest of the library goes on (exit code 1 if any failed).
|
|
32
|
+
|
|
33
|
+
## Selecting residues
|
|
34
|
+
|
|
35
|
+
`-r` / `residues=` takes chains and residue numbers, comma- or space-separated:
|
|
36
|
+
|
|
37
|
+
| selection | means |
|
|
38
|
+
|---|---|
|
|
39
|
+
| `A:2-10,B:20` | residues 2 to 10 of chain A, 20 of chain B |
|
|
40
|
+
| `A:2,4,6` | three residues of chain A (a chain holds until the next one) |
|
|
41
|
+
| `A:ARG239,THR315` | with names, which must then match |
|
|
42
|
+
| `100` / `100A` | any chain; `100` takes its insertion codes, `100A` only that one |
|
|
43
|
+
|
|
44
|
+
Residues with nothing to rotate (Gly, Ala) and Pro stay rigid, with a warning.
|
|
45
|
+
|
|
46
|
+
## Python API
|
|
47
|
+
|
|
48
|
+
```python
|
|
49
|
+
from rdkit import Chem
|
|
50
|
+
from rdkit.Chem import AllChem
|
|
51
|
+
from ms_adtools import prepare_ligand, prepare_receptor, prepare_flexreceptor, pdbqt_to_rdkit
|
|
52
|
+
|
|
53
|
+
mol = Chem.AddHs(Chem.MolFromSmiles("CC(=O)Nc1ccc(O)cc1"))
|
|
54
|
+
AllChem.EmbedMolecule(mol)
|
|
55
|
+
pdbqt = prepare_ligand(mol) # str, with REMARK SMILES / IDX / H PARENT
|
|
56
|
+
receptor = prepare_receptor("1iep.pdb") # str
|
|
57
|
+
rigid, flex = prepare_flexreceptor("1iep.pdbqt", "A:239,315")
|
|
58
|
+
poses = pdbqt_to_rdkit(docked_text) # one conformer per MODEL, `score` on each
|
|
59
|
+
```
|
|
60
|
+
|
|
61
|
+
| function | returns |
|
|
62
|
+
|---|---|
|
|
63
|
+
| `prepare_ligand(mol, open_macrocycles=False, hydrate=False, flexible_amides=False, ...)` | ligand PDBQT text |
|
|
64
|
+
| `prepare_reactive_ligand(mol, smarts, idx=0, ...)` | ligand PDBQT with the reactive atom typed |
|
|
65
|
+
| `prepare_receptor(path, repairs="", charges=None, cleanup=..., zinc=False, ccd=False)` | receptor PDBQT text |
|
|
66
|
+
| `prepare_flexreceptor(path, residues, reactive=())` | `(rigid, flex)` PDBQT texts |
|
|
67
|
+
| `pdbqt_to_rdkit(text)` | RDKit molecule, one conformer per pose |
|
|
68
|
+
| `dlg_to_pdbqt(text, cluster_leads=True)` | PDBQT poses of an AutoDock4/AutoDock-GPU DLG |
|
|
69
|
+
| `flexres_models(text, rigid=None)` | docked flexible residues as a multi-model PDB |
|
|
70
|
+
| `add_zinc_pseudo_atoms(pdbqt)` | receptor PDBQT with AutoDock4Zn TZ atoms |
|
|
71
|
+
|
|
72
|
+
`ms_adtools.reactive.reactive_config` writes the config of reactive docking;
|
|
73
|
+
`ms_adtools.zinc.PARAMETERS` is the `AD4Zn.dat` that autogrid4 needs for AutoDock4Zn maps.
|
|
74
|
+
Each function's docstring has the details.
|
|
75
|
+
|
|
76
|
+
## Speed and memory
|
|
77
|
+
|
|
78
|
+
One CPU, Python 3.12, RDKit 2026.09; each row is a fresh process.
|
|
79
|
+
|
|
80
|
+
| | wall | peak RSS |
|
|
81
|
+
|---|---|---|
|
|
82
|
+
| import | 0.12 s | 71 MB |
|
|
83
|
+
| 200 ligands (Enamine, 3D) | 0.29 s (1.5 ms each) | 72 MB |
|
|
84
|
+
| receptor (4uwh, 4905 atoms) | 0.16 s | 83 MB |
|
|
85
|
+
|
|
86
|
+
On `1iep_receptorH.pdb` the receptor has AutoDockTools' atoms and types (charges within 0.02 on
|
|
87
|
+
average), and `prepare_flexreceptor` writes its rigid file byte for byte. `benchmarks/` has the
|
|
88
|
+
script and the full measurements.
|
|
89
|
+
|
|
90
|
+
## License
|
|
91
|
+
|
|
92
|
+
MIT (`LICENSE`). The PDBQT conventions are AutoDockTools' (MGLTools, M. Sanner, TSRI) and the Forli
|
|
93
|
+
lab's; cite them with AutoDock. `src/ms_adtools/zinc/` carries AutoDock Vina's
|
|
94
|
+
`zinc_pseudo.py` (Apache-2.0) and the unmodified `AD4Zn.dat` parameter file (GPL-2.0-or-later
|
|
95
|
+
data), each with its own license (see its `NOTICE`).
|
|
@@ -0,0 +1,72 @@
|
|
|
1
|
+
# ms_adtools benchmark
|
|
2
|
+
|
|
3
|
+
Measured 2026-11-08 and 2026-11-10 on one machine: Python 3.12, RDKit 2026.09.1, numpy 2.5.4,
|
|
4
|
+
gemmi 0.7.6. meeko 0.8.0, the reference tool for AutoDock preparation, is measured alongside on
|
|
5
|
+
the same inputs. Reproduce with `python benchmarks/bench.py`; each row runs in a fresh
|
|
6
|
+
subprocess, so RSS is a clean-interpreter number.
|
|
7
|
+
|
|
8
|
+
Ligands: 200 Enamine_D10k molecules, hydrogens added, ETKDG + MMFF.
|
|
9
|
+
Receptor: `4uwh.pdb` (4905 atoms, `BENCH_PDB=.../4uwh.pdb`), read from the PDB file by both
|
|
10
|
+
(meeko with `allow_bad_res=True`, `default_altloc="A"`).
|
|
11
|
+
|
|
12
|
+
## Time and memory
|
|
13
|
+
|
|
14
|
+
All single-threaded.
|
|
15
|
+
|
|
16
|
+
| measurement | wall s | RSS growth MB | peak RSS MB |
|
|
17
|
+
|---|---|---|---|
|
|
18
|
+
| `from rdkit import Chem` alone | 0.03 | 30 | 42 |
|
|
19
|
+
| import ms_adtools, ligand path | 0.12 | 44 | 56 |
|
|
20
|
+
| import ms_adtools, receptor path | 0.12 | 59 | 71 |
|
|
21
|
+
| import meeko | 1.31 | 231 | 243 |
|
|
22
|
+
| 200 ligands, ms_adtools | 0.29 | 2 | 72 |
|
|
23
|
+
| 200 ligands, meeko | 0.43 | 1 | 268 |
|
|
24
|
+
| receptor, ms_adtools | 0.16 | 12 | 83 |
|
|
25
|
+
| receptor, meeko (`Polymer.from_pdb_file`) | 2.6 | 110 | 354 |
|
|
26
|
+
|
|
27
|
+
Opening macrocycles takes the 200 ligands to 0.31 s and hydration to 0.36 s. The number that
|
|
28
|
+
mattered for us is the per-worker footprint: a worker that only prepares receptors peaks at
|
|
29
|
+
about 80 MB.
|
|
30
|
+
|
|
31
|
+
## Output
|
|
32
|
+
|
|
33
|
+
`prepare_ligand` follows meeko's ligand conventions; over the same 200 ligands, in each mode,
|
|
34
|
+
its PDBQT matches meeko's:
|
|
35
|
+
|
|
36
|
+
| | rigid rings | open macrocycles | hydrated |
|
|
37
|
+
|---|---|---|---|
|
|
38
|
+
| same atoms written | 200/200 | 200/200 | 200/200 |
|
|
39
|
+
| same atom types | 200/200 | 189/200 | 200/200 |
|
|
40
|
+
| max charge difference | 0.000 | 0.000 | 0.000 |
|
|
41
|
+
| same TORSDOF | 200/200 | 200/200 | 200/200 |
|
|
42
|
+
| same number of BRANCHes | 200/200 | 200/200 | 200/200 |
|
|
43
|
+
| same W / G / CG counts | 200/200 | 200/200 | 200/200 |
|
|
44
|
+
| pose back to the input SMILES | 200/200 | 200/200 | 200/200 |
|
|
45
|
+
|
|
46
|
+
The 11 differences with open macrocycles are the choice of which C–C bond of the ring is cut
|
|
47
|
+
when several cuts give an equally shallow tree: the same number of breaks and the same TORSDOF,
|
|
48
|
+
one bond over.
|
|
49
|
+
|
|
50
|
+
The receptor follows AutoDockTools' conventions: bonds by distance (`DetermineConnectivity`),
|
|
51
|
+
bond orders and formal charges from the hydrogens the structure carries, Gasteiger over the
|
|
52
|
+
whole receptor. Against `prepare_receptor4.py` on `1iep_receptorH.pdb` it writes the same 2702
|
|
53
|
+
atoms with the same types, charges within 0.020 on average (ADT charges residue by residue);
|
|
54
|
+
`prepare_flexreceptor` writes ADT's rigid file byte for byte and the same torsion tree.
|
|
55
|
+
|
|
56
|
+
## Scope
|
|
57
|
+
|
|
58
|
+
What each covers, for choosing between them:
|
|
59
|
+
|
|
60
|
+
| | ms_adtools | meeko |
|
|
61
|
+
|---|---|---|
|
|
62
|
+
| ligand PDBQT, `REMARK SMILES` / `IDX` / `H PARENT` | yes | yes |
|
|
63
|
+
| pose back to an RDKit molecule | yes | yes |
|
|
64
|
+
| receptor PDBQT | yes, by distances and hydrogens; CCD templates for non-standard residues | yes, by residue templates |
|
|
65
|
+
| flexible residues | yes | yes |
|
|
66
|
+
| macrocycle opening, hydrated docking, flexible amides | yes | yes |
|
|
67
|
+
| reactive docking | yes | yes |
|
|
68
|
+
| covalent docking | no | yes |
|
|
69
|
+
| AutoDock4 DLG poses | yes | yes |
|
|
70
|
+
| AutoDock4Zn TZ pseudo-atoms | yes | no |
|
|
71
|
+
| receptor input | PDB, mmCIF, PQR, PDBQT, mol2 | PDB, PQR, mmCIF, prody selection |
|
|
72
|
+
| required imports | rdkit, numpy, gemmi | rdkit, numpy, scipy, gemmi, pandas, prody |
|
|
@@ -0,0 +1,128 @@
|
|
|
1
|
+
"""ms_adtools benchmark, meeko as the reference: import cost, ligand and receptor preparation time and memory.
|
|
2
|
+
|
|
3
|
+
Each measurement runs in its own subprocess, so the reported RSS is that of a fresh
|
|
4
|
+
interpreter that only imported what the measurement needs.
|
|
5
|
+
|
|
6
|
+
python bench.py # all measurements, as a markdown table
|
|
7
|
+
python bench.py import_meeko # one measurement, as JSON
|
|
8
|
+
"""
|
|
9
|
+
import json
|
|
10
|
+
import os
|
|
11
|
+
import subprocess
|
|
12
|
+
import sys
|
|
13
|
+
import time
|
|
14
|
+
|
|
15
|
+
HERE = os.path.dirname(os.path.abspath(__file__))
|
|
16
|
+
LIGANDS = os.environ.get("BENCH_SDF", "/tmp/adt/ligs/set200.sdf")
|
|
17
|
+
RECEPTOR = os.environ.get("BENCH_PDB", os.path.join(HERE, "..", "test", "data", "1iep_receptorH.pdb"))
|
|
18
|
+
|
|
19
|
+
|
|
20
|
+
def rss_mb():
|
|
21
|
+
with open("/proc/self/status") as f:
|
|
22
|
+
for line in f:
|
|
23
|
+
if line.startswith("VmRSS:"):
|
|
24
|
+
return int(line.split()[1]) / 1024
|
|
25
|
+
return float("nan")
|
|
26
|
+
|
|
27
|
+
|
|
28
|
+
def _timed(fn):
|
|
29
|
+
"""(wall seconds, cpu seconds, RSS growth MB, peak RSS MB) of fn()."""
|
|
30
|
+
base = rss_mb()
|
|
31
|
+
t, c = time.perf_counter(), time.process_time()
|
|
32
|
+
fn()
|
|
33
|
+
wall, cpu = time.perf_counter() - t, time.process_time() - c
|
|
34
|
+
return wall, cpu, rss_mb() - base, rss_mb()
|
|
35
|
+
|
|
36
|
+
|
|
37
|
+
def import_adtools_receptor():
|
|
38
|
+
def go():
|
|
39
|
+
from ms_adtools import prepare_receptor # noqa: F401 (numpy only)
|
|
40
|
+
return _timed(go)
|
|
41
|
+
|
|
42
|
+
|
|
43
|
+
def import_adtools_ligand():
|
|
44
|
+
def go():
|
|
45
|
+
from ms_adtools import prepare_ligand # noqa: F401 (pulls RDKit in)
|
|
46
|
+
return _timed(go)
|
|
47
|
+
|
|
48
|
+
|
|
49
|
+
def import_rdkit():
|
|
50
|
+
def go():
|
|
51
|
+
from rdkit import Chem # noqa: F401 (the floor both ligand paths pay)
|
|
52
|
+
return _timed(go)
|
|
53
|
+
|
|
54
|
+
|
|
55
|
+
def import_meeko():
|
|
56
|
+
def go():
|
|
57
|
+
import meeko # noqa: F401
|
|
58
|
+
return _timed(go)
|
|
59
|
+
|
|
60
|
+
|
|
61
|
+
def _mols():
|
|
62
|
+
from rdkit import Chem
|
|
63
|
+
return [m for m in Chem.SDMolSupplier(LIGANDS, removeHs=False) if m is not None]
|
|
64
|
+
|
|
65
|
+
|
|
66
|
+
def ligands_adtools():
|
|
67
|
+
from ms_adtools import prepare_ligand
|
|
68
|
+
mols = _mols()
|
|
69
|
+
out = []
|
|
70
|
+
return (*_timed(lambda: out.extend(prepare_ligand(m) for m in mols)), len(mols))
|
|
71
|
+
|
|
72
|
+
|
|
73
|
+
def ligands_meeko():
|
|
74
|
+
from rdkit import Chem
|
|
75
|
+
from meeko import MoleculePreparation, PDBQTWriterLegacy
|
|
76
|
+
# meeko refuses a molecule with more than one fragment, so the salts are split here,
|
|
77
|
+
# as ms_adtools does on its own
|
|
78
|
+
mols = [max(Chem.GetMolFrags(m, asMols=True), key=lambda f: f.GetNumHeavyAtoms()) for m in _mols()]
|
|
79
|
+
prep = MoleculePreparation()
|
|
80
|
+
out = []
|
|
81
|
+
|
|
82
|
+
def go():
|
|
83
|
+
for m in mols:
|
|
84
|
+
for setup in prep.prepare(m):
|
|
85
|
+
out.append(PDBQTWriterLegacy.write_string(setup)[0])
|
|
86
|
+
return (*_timed(go), len(mols))
|
|
87
|
+
|
|
88
|
+
|
|
89
|
+
def receptor_adtools():
|
|
90
|
+
from ms_adtools import prepare_receptor
|
|
91
|
+
return _timed(lambda: prepare_receptor(RECEPTOR))
|
|
92
|
+
|
|
93
|
+
|
|
94
|
+
def receptor_meeko():
|
|
95
|
+
from meeko import MoleculePreparation, Polymer, ResidueChemTemplates
|
|
96
|
+
|
|
97
|
+
def go():
|
|
98
|
+
Polymer.from_pdb_file(RECEPTOR, ResidueChemTemplates.create_from_defaults(),
|
|
99
|
+
MoleculePreparation(), blunt_ends=[(0, 0)], allow_bad_res=True,
|
|
100
|
+
default_altloc="A")
|
|
101
|
+
return _timed(go)
|
|
102
|
+
|
|
103
|
+
|
|
104
|
+
MEASUREMENTS = {k: v for k, v in list(globals().items())
|
|
105
|
+
if k.startswith(("import_", "ligands_", "receptor_"))}
|
|
106
|
+
|
|
107
|
+
|
|
108
|
+
def main():
|
|
109
|
+
if len(sys.argv) > 1:
|
|
110
|
+
print(json.dumps(MEASUREMENTS[sys.argv[1]]()))
|
|
111
|
+
return
|
|
112
|
+
rows = {}
|
|
113
|
+
for name in MEASUREMENTS:
|
|
114
|
+
p = subprocess.run([sys.executable, __file__, name], capture_output=True, text=True)
|
|
115
|
+
last = p.stdout.strip().splitlines()[-1:] # the tools print notes of their own
|
|
116
|
+
rows[name] = json.loads(last[0]) if p.returncode == 0 and last else p.stderr.strip().splitlines()[-1:]
|
|
117
|
+
print("| measurement | wall s | cpu s | RSS growth MB | peak RSS MB | n |")
|
|
118
|
+
print("|---|---|---|---|---|---|")
|
|
119
|
+
for name, r in rows.items():
|
|
120
|
+
if isinstance(r, list) and len(r) >= 4:
|
|
121
|
+
n = r[4] if len(r) > 4 else ""
|
|
122
|
+
print(f"| {name} | {r[0]:.2f} | {r[1]:.2f} | {r[2]:.0f} | {r[3]:.0f} | {n} |")
|
|
123
|
+
else:
|
|
124
|
+
print(f"| {name} | failed: {r} | | | | |")
|
|
125
|
+
|
|
126
|
+
|
|
127
|
+
if __name__ == "__main__":
|
|
128
|
+
main()
|